with whom Justice Blackmun joins, dissenting.
Power, not reason, is the new currency of this Court’s deci-sionmaking. Four Terms ago, a five-Justice majority of this Court held that “victim impact” evidence of the type at issue in this case could not constitutionally be introduced during the penalty phase of a capital trial. Booth v. Maryland, 482 U. S. 496 (1987). By another 5-4 vote, a majority of this Court rebuffed an attack upon this ruling just two Terms ago. South Carolina v. Gathers, 490 U. S. 805 (1989). Nevertheless, having expressly invited respondent to renew the attack, 498 U. S. 1076 (1991), today’s majority overrules Booth and Gathers and credits the dissenting views expressed in those cases. Neither the law nor the facts supporting Booth and Gathers underwent any change in the last four years. Only the personnel of this Court did.
In dispatching Booth and Gathers to their graves, today’s majority ominously suggests that an even more extensive upheaval of this Court’s precedents may be in store. Renouncing this Court’s historical commitment to a conception of “the judiciary as a source of impersonal and reasoned judgments,” Moragne v. States Marine Lines, 398 U. S. 375, 403 (1970), *845the majority declares itself free to discard any principle of constitutional liberty which was recognized or reaffirmed over the dissenting votes of four Justices and with which five or more Justices now disagree. The implications of this radical new exception to the doctrine of stare decisis are staggering. The majority today sends a clear signal that scores of established constitutional liberties are now ripe for reconsideration, thereby inviting the very type of open defiance of our precedents that the majority rewards in this case. Because I believe that this Court owes more to its constitutional precedents in general and to Booth and Gathers in particular, I dissent.
I
Speaking for the Court as then constituted, Justice Powell and Justice Brennan set out the rationale for excluding victim-impact evidence from the sentencing proceedings in a capital case. See Booth v. Maryland, supra, at 504-509; South Carolina v. Gathers, supra, at 810-811. As the majorities in Booth and Gathers recognized, the core principle of this Court’s capital jurisprudence is that the sentence of death must reflect an “‘individualized determination’” of the defendant’s “‘personal responsibility and moral guilt’” and must be based upon factors that channel the jury’s discretion “ ‘so as to minimize the risk of wholly arbitrary and capricious action.’” Booth v. Maryland, supra, at 502, quoting Zant v. Stephens, 462 U. S. 862, 879 (1983); Enmund v. Florida, 458 U. S. 782, 801 (1982), and Gregg v. Georgia, 428 U. S. 153, 189 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.); accord, South Carolina v. Gathers, supra, at 810. The State’s introduction of victim-impact evidence, Justice Powell and Justice Brennan explained, violates this fundamental principle. Where, as is ordinarily the case, the defendant was unaware of the personal circumstances of his victim, admitting evidence of the victim’s character and the impact of the murder upon the victim’s family predicates the sentencing determination on “factors . . . wholly unrelated to the *846blameworthiness of [the] particular defendant.” Booth v. Maryland, supra, at 504; South Carolina v. Gathers, supra, at 810. And even where the defendant was in a position to foresee the likely impact of his conduct, admission of victim-impact evidence creates an unacceptable risk of sentencing arbitrariness. As Justice Powell explained in Booth, the probative value of such evidence is always outweighed by its prejudicial effect because of its inherent capacity to draw the jury’s attention away from the character of the defendant and the circumstances of the crime to such illicit considerations as the eloquence with which family members express their grief and the status of the victim in the community. See Booth v. Maryland, supra, at 505-507, and n. 8; South Carolina v. Gathers, supra, at 810-811. I continue to find these considerations wholly persuasive, and I see no purpose in trying to improve upon Justice Powell’s and Justice Brennan’s exposition of them.
There is nothing new in the majority’s discussion of the supposed deficiencies in Booth and Gathers. Every one of the arguments made by the majority can be found in the dissenting opinions filed in those two cases, and, as I show in the margin, each argument was convincingly answered by Justice Powell and Justice Brennan.1
*847But contrary to the impression that one might receive from reading the majority’s lengthy rehearsing of the issues addressed in Booth and Gathers, the outcome of this case does *848not turn simply on who — the Booth and Gathers majorities or the Booth and Gathers dissenters — had the better of the argument. Justice Powell and Justice Brennan’s position carried the day in those cases and became the law of the land. The real question, then, is whether today’s majority has come forward with the type of extraordinary showing that this Court has historically demanded before overruling one of its precedents. In my view, the majority clearly has not made any such showing. Indeed, the striking feature of the majority’s opinion is its radical assertion that it need not even try.
II
The overruling of one of this Court’s precedents ought to be a matter of great moment and consequence. Although the doctrine of stare decisis is not an “inexorable command,” Burnet v. Coronado Oil & Gas Co., 285 U. S. 393, 405 (1932) (Brandeis, J., dissenting), this Court has repeatedly stressed that fidelity to precedent is fundamental to “a society governed by the rule of law,” Akron v. Akron Center for Reproductive Health, Inc., 462 U. S. 416, 420 (1983). See generally Patterson v. McLean Credit Union, 491 U. S. 164, 172 (1989) (“[I]t is indisputable that stare decisis is a basic self-governing principle within the Judicial Branch, which is entrusted with the sensitive and difficult task of fashioning and preserving a jurisprudential system that is not based upon *849‘an arbitrary discretion.’ The Federalist, No. 78, p. 490 (H. Lodge ed. 1888) (A. Hamilton)”); Appeal of Concerned Corporators of Portsmouth Savings Bank, 129 N. H. 183, 227, 525 A. 2d 671, 701 (1987) (Souter, J., dissenting) (“[S]tare decisis ... ‘is essential if case-by-case judicial decision-making is to be reconciled with the principle of the rule of law, for when governing legal standards are open to revision in every case, deciding cases becomes a mere exercise of judicial will, with arbitrary and unpredictable results,”’ quoting Thornburgh v. American College of Obstetricians and Gynecologists, 476 U. S. 747, 786-787 (1986) (White, J., dissenting)).
Consequently, this Court has never departed from precedent without “special justification.” Arizona v. Rumsey, 467 U. S. 203, 212 (1984). Such justifications include the advent of “subsequent changes or development in the law” that undermine a decision’s rationale, Patterson v. McLean Credit Union, supra, at 173; the need “to bring [a decision] into agreement with experience and with facts newly ascertained,” Burnet v. Coronado Oil & Gas Co., supra, at 412 (Brandeis, J., dissenting); and a showing that a particular precedent has become a “detriment to coherence and consistency in the law,” Patterson v. McLean Credit Union, supra, at 173.
The majority cannot seriously claim that any of these traditional bases for overruling a precedent applies to Booth or Gathers. The majority does not suggest that the legal rationale of these decisions has been undercut by changes or developments in doctrine during the last two years. Nor does the majority claim that experience over that period of time has discredited the principle that “any decision to impose the death sentence be, and appear to be, based on reason rather than caprice or emotion,” Gardner v. Florida, 430 U. S. 349, 358 (1977) (plurality opinion), the larger postulate of political morality on which Booth and Gathers rest.
The majority does assert that Booth and Gathers “have defied consistent application by the lower courts,” ante, at 830, *850but the evidence that the majority proffers is so feeble that the majority cannot sincerely expect anyone to believe this claim. To support its contention, the majority points to Justice O’Connor’s dissent in Gathers, which noted a division among lower courts over whether Booth prohibited prosecutorial arguments relating to the victim’s personal characteristics. See 490 U. S., at 813. That, of course, was the issue expressly considered and resolved in Gathers. The majority also cites The Chief Justice’s dissent in Mills v. Maryland, 486 U. S. 367, 395-398 (1988). That opinion does not contain a single word about any supposed “[in]consistent application” of Booth in the lower courts. Finally, the majority refers to a divided Ohio Supreme Court decision disposing of an issue concerning victim-impact evidence. See State v. Huertas, 51 Ohio St. 3d 22, 553 N. E. 2d 1058 (1990), cert. dism’d as improvidently granted, 498 U. S. 336 (1991). Obviously, if a division among the members of a single lower court in a single case were sufficient to demonstrate that a particular precedent was a “detriment to coherence and consistency in the law,” Patterson v. McLean Credit Union, supra, at 173, there would hardly be a decision in United States Reports that we would not be obliged to reconsider.
It takes little real detective work to discern just what has changed since this Court decided Booth and Gathers: this Court’s own personnel. Indeed, the majority candidly explains why this particular contingency, which until now has been almost universally understood not to be sufficient to warrant overruling a precedent, see, e. g., Florida Dept. of Health and Rehabilitative Services v. Florida Nursing Home Assn., 450 U. S. 147, 153 (1981) (Stevens, J., concurring); Mitchell v. W. T. Grant Co., 416 U. S. 600, 636 (1974) (Stewart, J., dissenting); Mapp v. Ohio, 367 U. S. 643, 677 (1961) (Harlan, J., dissenting); but see South Carolina v. Gathers, supra, at 824 (Scalia, J., dissenting), is sufficient to justify overruling Booth and Gathers. “Considerations in favor of stare decisis are at their acme,” the majority explains, “in *851cases involving property and contract rights, where reliance interests are involved[;] the opposite is true in cases such as the present one involving procedural and evidentiary rules.” Ante, at 828 (citations omitted). In addition, the majority points out, “Booth and Gathers were decided by the narrowest of margins, over spirited dissents” and thereafter were “questioned by Members of the Court.” Ante, at 828-829. Taken together, these considerations make it legitimate, in the majority’s view, to elevate the position of the Booth and Gathers dissenters into the law of the land.
This truncation of the Court’s duty to stand by its own precedents is astonishing. By limiting full protection of the doctrine of stare decisis to “cases involving property and contract rights,” ante, at 828, the majority sends a clear signal that essentially all decisions implementing the personal liberties protected by the Bill of Rights and the Fourteenth Amendment are open to reexamination. Taking into account the majority’s additional criterion for overruling — that a case either was decided or reaffirmed by a 5-4 margin “over spirited dissen[t],” ante, at 829 — the continued vitality of literally scores of decisions must be understood to depend on nothing more than the proclivities of the individuals who now comprise a majority of this Court. See, e. g., Metro Broadcasting v. FCC, 497 U. S. 547 (1990) (authority of Federal government to set aside broadcast licenses for minority applicants); Grady v. Corbin, 495 U. S. 508 (1990) (right under Double Jeopardy Clause not to be subjected twice to prosecution for same criminal conduct); Mills v. Maryland, supra (Eighth Amendment right to jury instructions that do not preclude consideration of nonunanimous mitigating factors in capital sentencing); United States v. Paradise, 480 U. S. 149 (1987) (right to promotions as remedy for racial discrimination in government hiring); Ford v. Wainwright, 477 U. S. 399 (1986) (Eighth Amendment right not to be executed if insane); Thornburgh v. American College of Obstetricians and Gynecologists, 476 U. S. 747 (1986) (reaffirming *852right to abortion recognized in Roe v. Wade, 410 U. S. 113 (1973)); Aguilar v. Felton, 473 U. S. 402 (1985) (Establishment Clause bar on governmental financial assistance to parochial schools).2
In my view, this impoverished conception of stare decisis cannot possibly be reconciled with the values that inform the proper judicial function. Contrary to what the majority suggests, stare decisis is important not merely because individuals rely on precedent to structure their commercial activity but because fidelity to precedent is part and parcel of a conception of “the judiciary as a source of impersonal and reasoned judgments.” Moragne v. States Marine Lines, 398 U. S., at 403. Indeed, this function of stare decisis is in many respects even more critical in adjudication involving constitutional liberties than in adjudication involving com*853mercial entitlements. Because enforcement of the Bill of Rights and the Fourteenth Amendment frequently requires this Court to rein in the forces of democratic politics, this Court can legitimately lay claim to compliance with its directives only if the public understands the Court to be implementing “principles . . . founded in the law rather than in the proclivities of individuals.” Vasquez v. Hillery, 474 U. S. 254, 265 (1986).3 Thus, as Justice Stevens has explained, the “stron[g] presumption of validity” to which “recently decided cases” are entitled “is an essential thread in the mantle of protection that the law affords the individual. ... It is the unpopular or beleaguered individual — not the man in power — who has the greatest stake in the integrity of the law.” Florida Dept. of Health and Rehabilitative Services v. Florida Nursing Home Assn., 450 U. S., at 153-154 (concurring opinion).
Carried to its logical conclusion, the majority’s debilitated conception of stare decisis would destroy the Court’s very capacity to resolve authoritatively the abiding conflicts between those with power and those without. If this Court shows so little respect for its own precedents, it can hardly expect them to be treated more respectfully by the state actors whom these decisions are supposed to bind. See *854Mitchell v. W. T. Grant Co., 416 U. S., at 634 (Stewart, J., dissenting). By signaling its willingness to give fresh consideration to any constitutional liberty recognized by a 5-4 vote “over spirited dissen[t],” ante, at 829, the majority invites state actors to renew the very policies deemed unconstitutional in the hope that this Court may now reverse course, even if it has only recently reaffirmed the constitutional liberty in question.
Indeed, the majority’s disposition of this case nicely illustrates the rewards of such a strategy of defiance. The Tennessee Supreme Court did nothing in this case to disguise its contempt for this Court’s decisions in Booth and Gathers. Summing up its reaction to those cases, it concluded:
“It is an affront to the civilized members of the human race to say that at sentencing in a capital case, a parade of witnesses may praise the background, character and good deeds of Defendant (as was done in this case), without limitation as to relevancy, but nothing may be said that bears upon the character of, or harm imposed, upon the victims.” 791 S. W. 2d 10, 19 (1990).
Offering no explanation for how this case could possibly be distinguished from Booth and Gathers — for obviously, there is none to offer — the court perfunctorily declared that the victim-impact evidence and the prosecutor’s argument based on this evidence “did not violate either [of those decisions].” Ibid. It cannot be clearer that the court simply declined to be bound by this Court’s precedents.4
*855Far from condemning this blatant disregard for the rule of law, the majority applauds it. In the Tennessee Supreme Court’s denigration of Booth and Gathers as “‘an affront to the civilized members of the human race,’ ” the majority finds only confirmation of “the unfairness of the rule pronounced by” the majorities in those cases. Ante, at 826. It is hard to imagine a more complete abdication of this Court’s historic commitment to defending the supremacy of its own pronouncements on issues of constitutional liberty. See Cooper v. Aaron, 358 U. S. 1 (1958); see also Hutto v. Davis, 454 U. S. 370, 375 (1982) (per curiam) (“[U]nless we wish anarchy to prevail within the federal judicial system, a precedent of this Court must be followed by the lower federal courts no matter how misguided the judges of those courts may think it to be”). In light of the cost that such abdication exacts on the authoritativeness of all of this Court’s pronouncements, it is also hard to imagine a more short-sighted strategy for effecting change in our constitutional order.
*856rH h — i h — I
Today’s decision charts an unmistakable course. If the majority’s radical reconstruction of the rules for overturning this Court’s decisions is to be taken at face value — and the majority offers us no reason why it should not — then the overruling of Booth and Gathers is but a preview of an even broader and more far-reaching assault upon this Court’s precedents. Cast aside today are those condemned to face society’s ultimate penalty. Tomorrow’s victims may be minorities, women, or the indigent. Inevitably, this campaign to resurrect yesterday’s “spirited dissents” will squander the authority and the legitimacy of this Court as a protector of the powerless.
I dissent.
The majority’s primary argument is that punishment in criminal law is frequently based on an “assessment of [the] harm caused by the defendant as a result of the crime charged.” Ante, at 819. See also Booth v. Maryland, 482 U. S. 496, 616 (1987) (White, J., dissenting); id., at 619-520 (SCALIA, J., dissenting); South Carolina v. Gathers, 490 U. S. 805, 818-819 (1989) (O’Connor, J., dissenting). Nothing in Booth or Gathers, however, conflicts with this unremarkable observation. These cases stand merely for the proposition that the State may not put on evidence of one particular species of harm — namely, that associated with the victim’s personal characteristics independent of the circumstances of the offense— in the course of a capital murder proceeding. See Booth v. Maryland, supra, at 507, n. 10 (emphasizing that decision does not bar reliance on victim-impact evidence in capital sentencing so long as such evidence “relate^] directly to the circumstances of the crime”); id., at 509, n. 12 (emphasizing that decision does not bar reliance on victim-impact evidence *847in sentencing for noncapital crimes). It may be the case that such a rule departs from the latitude of sentencers in criminal law generally to “tak[e] into consideration the harm done by the defendant.” Ante, at 825. But as the Booth Court pointed out, because this Court’s capital-sentencing jurisprudence is founded on the premise that “death is a ‘punishment different from all other sanctions,”’ it is completely unavailing to attempt to infer from sentencing considerations in noncapital settings the proper treatment of any particular sentencing issue in a capital case. 482 U. S., at 509, n. 12, quoting Woodson v. North Carolina, 428 U. S. 280, 303-304, 305 (1976) (opinion of Stewart, Powell, and Stevens, JJ.).
The majority also discounts Justice Powell’s concern with the inherently prejudicial quality of victim-impact evidence. “[T]he mere fact that for tactical reasons it might not be prudent for the defense to rebut victim impact evidence,” the majority protests, “makes the case no different than others in which a party is faced with this sort of a dilemma.” Ante, at 823. See also Booth v. Maryland, supra, at 518 (White, J., dissenting). Unsurprisingly, this tautology is completely unresponsive to Justice Powell’s argument. The Booth Court established a rule excluding introduction of victim-impact evidence not merely because it is difficult to rebut — a feature of victim-impact evidence that may be “no different” from that of many varieties of relevant, legitimate evidence — but because the effect of this evidence in the sentencing proceeding is unfairly prejudicial: “The prospect of a ‘mini-trial’ on the victim’s character is more than simply unappealing; it could well distract the sentencing jury from its constitutionally required task — determining whether the death penalty is appropriate in light of the background and record of the accused and the particular circumstances of the crime.” 482 U. S., at 507. The law is replete with per se prohibitions of types of evidence the probative effect of which is generally outweighed by its unfair prejudice. See, e. g., Fed. Rules Evid. 404, 407-412. There is nothing anomalous in the notion that the Eighth Amendment would similarly exclude evidence that has an undue capacity to undermine the regime of individualized sentencing that our capital jurisprudence demands.
Finally, the majority contends that the exclusion of victim-impact evidence “deprives the State of the full moral force of its evidence and may prevent the jury from having before it all the information necessary to determine the proper punishment for a first-degree murder.” Ante, at 825. The majority’s recycled contention, see Booth, supra, at 517 (White, J., *848dissenting); id., at 520 (Scalia, J., dissenting); Gathers, supra, at 817-818 (O'Connor, J., dissenting), begs the question. Before it is possible to conclude that the exclusion of victim-impact evidence prevents the State from making its case or the jury from considering relevant evidence, it is necessary to determine whether victim-impact evidence is consistent with the substantive standards that define the scope of permissible sentencing determinations under the Eighth Amendment. The majority offers no persuasive answer to Justice Powell and Justice Brennan’s conclusion that victim-impact evidence is frequently irrelevant to any permissible sentencing consideration and that such evidence risks exerting illegitimate “moral force” by directing the jury’s attention on illicit considerations such as the victim’s standing in the community.
Based on the majority’s new criteria for overruling, these decisions, too, must be included on the “endangered precedents” list: Rutan v. Republican Party of Illinois, 497 U. S. 62 (1990) (First Amendment right not to be denied public employment on the basis of party affiliation); Peel v. Attorney Registration and Disciplinary Comm’n of Ill., 496 U. S. 91 (1990) (First Amendment right to advertise legal specialization); Zinermon v. Burch, 494 U. S. 113 (1990) (due process right to procedural safeguards aimed at assuring voluntariness of decision to commit oneself to mental hospital); James v. Illinois, 493 U. S. 307 (1990) (Fourth Amendment right to exclusion of illegally obtained evidence introduced for impeachment of defense witness); Rankin v. McPherson, 483 U. S. 378 (1987) (First Amendment right of public employee to express views on matter of public importance); Rock v. Arkansas, 483 U. S. 44 (1987) (Fifth Amendment and Sixth Amendment right of criminal defendant to provide hypnotically refreshed testimony on his own behalf); Gray v. Mississippi, 481 U. S. 648 (1987) (rejecting applicability of harmless error analysis to Eighth Amendment right not to be sentenced to death by “death qualified” jury); Maine v. Moulton, 474 U. S. 159 (1985) (Sixth Amendment right to counsel violated by introduction of statements made to government informantcodefendant in course of preparing defense strategy); Garcia v. San Antonio Metropolitan Transit Authority, 469 U. S. 528 (1985) (rejecting theory that Tenth Amendment provides immunity to States from federal regulation); Pulliam v. Allen, 466 U. S. 522 (1984) (right to obtain injunctive relief from constitutional violations committed by judicial officials).
It does not answer this concern to suggest that Justices owe fidelity to the text of the Constitution rather than to the case law of this Court interpreting the Constitution. See, e. g., South Carolina v. Gathers, 490 U. S., at 825 (Scalia, J., dissenting). The text of the Constitution is rarely so plain as to be self-executing; invariably, this Court must develop mediating principles and doctrines in order to bring the text of constitutional provisions to bear on particular facts. Thus, to rebut the charge of personal lawmaking, Justices who would discard the mediating principles embodied in precedent must do more than state that they are following the “text” of the Constitution; they must explain why they are entitled to substitute their mediating principles for those that are already settled in the law. And such an explanation will be sufficient to legitimize the departure from precedent only if it measures up to the extraordinary standard necessary to justify overruling one of this Court’s precedents. See generally Note, 103 Harv. L. Rev. 1344, 1351-1354 (1990).
Equally unsatisfactory is the Tennessee Supreme Court’s purported finding that any error associated with the victim-impact evidence in this ease was harmless. See 791 S. W. 2d, at 19. This finding was based on the court’s conclusion that “the death penalty was the only rational punishment available” in light of the “inhuman brutality” evident in the circumstances of the murder. Ibid. It is well established that a State cannot make the death penalty mandatory for any class of aggravated murder; no matter how “brutal” the circumstances of the offense, the State must permit the sentencer discretion to impose a sentence of less than death. See *855Roberts v. Louisiana, 428 U. S. 325 (1976); Woodson v. North Carolina, 428 U. S. 280 (1976). It follows that an appellate court cannot deem error to be automatically harmless based solely on the aggravated character of a murder without assessing the impact of the error on the sentencer’s discretion. Cf. Clemons v. Mississippi, 494 U. S. 738, 751-752 (1990).
To sentence petitioner to death, the jury was required to find that the mitigating circumstances shown by petitioner did not outweigh the aggravating circumstances. See App. 21-22. In what it tried to pass off as harmless error analysis, the Tennessee Supreme Court failed to address how the victim-impact evidence introduced diming the sentencing proceedings in this ease likely affected the jury’s determination that the balance of aggravating and mitigating circumstances dictated a death sentence. Outside of a videotape of the crime scene, the State introduced no additional substantive evidence in the penalty phase other than the testimony of Mary Zvolanek, mother and grandmother of the murder victims. See 791 S. W. 2d, at 17. Under these circumstances, it is simply impossible to conclude that this victim-impact testimony, combined with the prosecutor’s extrapolation from it in his closing argument, was harmless beyond a reasonable doubt.