Deck v. Missouri, 544 U.S. 622 (2005)
OCTOBER TERM, 2004
DECK V. MISSOURI
SUPREME COURT OF THE UNITED STATES
DECK v. MISSOURI
certiorari to the supreme court of missouri
No. 04–5293.Argued March 1, 2005—Decided May 23, 2005
Petitioner Deck was convicted of capital murder and sentenced to death, but the Missouri Supreme Court set aside the sentence. At his new sentencing proceeding, he was shackled with leg irons, handcuffs, and a belly chain. The trial court overruled counsel’s objections to the shackles, and Deck was again sentenced to death. Affirming, the State Supreme Court rejected Deck’s claim that his shackling violated, inter alia, the Federal Constitution.
Held: The Constitution forbids the use of visible shackles during a capital trial’s penalty phase, as it does during the guilt phase, unless that use is “justified by an essential state interest”—such as courtroom security—specific to the defendant on trial. Holbrook v. Flynn, 475 U. S. 560, 568–569. Pp. 3–10.
(a) The law has long forbidden routine use of visible shackles during a capital trial’s guilt phase, permitting shackling only in the presence of a special need. In light of Holbrook, Illinois v. Allen, 397 U. S. 337, early English cases, and lower court shackling doctrine dating back to the 19th century, it is now clear that this is a basic element of due process protected by the Federal Constitution. Thus, the Fifth and Fourteenth Amendments prohibit using physical restraints visible to the jury absent a trial court determination, in the exercise of its discretion, that restraints are justified by a state interest specific to the particular defendant on trial. Pp. 3–6.
(b) If the reasons motivating the guilt phase constitutional rule—the presumption of innocence, securing a meaningful defense, and maintaining dignified proceedings—apply with like force at the penalty phase, the same rule will apply there. The latter two considerations obviously apply. As for the first, while the defendant’s conviction means that the presumption of innocence no longer applies, shackles at the penalty phase threaten related concerns. The jury, though no longer deciding between guilt and innocence, is deciding between life and death, which, given the sanction’s severity and finality, is no less important, Monge v. California, 524 U. S. 721, 732. Nor is accuracy in making that decision any less critical. Yet, the offender’s appearance in shackles almost inevitably implies to a jury that court authorities consider him a danger to the community (which is often a statutory aggravator and always a relevant factor); almost inevitably affects adversely the jury’s perception of the defendant’s character; and thereby inevitably undermines the jury’s ability to weigh accurately all relevant considerations when determining whether the defendant deserves death. The constitutional rule that courts cannot routinely place defendants in shackles or other restraints visible to the jury during the penalty phase is not absolute. In the judge’s discretion, account may be taken of special circumstances in the case at hand, including security concerns, that may call for shackling in order to accommodate the important need to protect the courtroom and its occupants. Pp. 6–10.
(c) Missouri’s arguments that its high court’s decision in this case meets the Constitution’s requirements are unconvincing. The first—that that court properly concluded that there was no evidence that the jury saw the restraints—is inconsistent with the record, which shows that the jury was aware of them, and overstates what the court actually said, which was that trial counsel made no record of the extent of the jury’s awareness of the shackles. The second—that the trial court acted within its discretion—founders on the record, which does not clearly indicate that the judge weighted the particular circumstances of the case. The judge did not refer to an escape risk or threat to courtroom security or explain why, if shackles were necessary, he did not provide nonvisible ones as was apparently done during the guilt phase of this case. The third—that Deck suffered no prejudice—fails to take account of Holbrook’s statement that shackling is “inherently prejudicial,” 475 U. S., at 568, a view rooted in this Court’s belief that the practice will often have negative effects that “cannot be shown from a trial transcript,” Riggins v. Nevada, 504 U. S. 127, 137. Thus, where a court, without adequate justification, orders the defendant to wear shackles visible to the jury, the defendant need not demonstrate actual prejudice to make out a due process violation. The State must prove “beyond a reasonable doubt that the [shackling] did not contribute to the verdict obtained.” Chapman v. California, 386 U. S. 18, 24. Pp. 10–12.
136 S. W. 3d 481, reversed and remanded.
Breyer, J., delivered the opinion of the Court, in which Rehnquist, C. J., and Stevens, O’Connor, Kennedy, Souter, and Ginsburg, JJ., joined. Thomas, J., filed a dissenting opinion, in which Scalia, J., joined.
OPINION OF THE COURT
DECK V. MISSOURI
544 U. S. ____ (2005)
SUPREME COURT OF THE UNITED STATES
CARMAN L. DECK, PETITIONER v. MISSOURI on writ of certiorari to the supreme court of missouri [May 23, 2005] Justice Breyer delivered the opinion of the Court. We here consider whether shackling a convicted offender during the penalty phase of a capital case violates the Federal Constitution. We hold that the Constitution forbids the use of visible shackles during the penalty phase, as it forbids their use during the guilt phase, unless that use is “justified by an essential state interest”—such as the interest in courtroom security—specific to the defendant on trial. Holbrook v. Flynn, 475 U. S. 560, 568–569 (1986); see also Illinois v. Allen, 397 U. S. 337, 343–344 (1970). I In July 1996, petitioner Carman Deck robbed, shot, and killed an elderly couple. In 1998, the State of Missouri tried Deck for the murders and the robbery. At trial, state authorities required Deck to wear leg braces that apparently were not visible to the jury. App. 5; Tr. of Oral Arg. 21, 25, 29. Deck was convicted and sentenced to death. The State Supreme Court upheld Deck’s conviction but set aside the sentence. 68 S. W. 3d 418, 432 (2002). The State then held a new sentencing proceeding. From the first day of the new proceeding, Deck was shackled with leg irons, handcuffs, and a belly chain. App. 58. Before the jury voir dire began, Deck’s counsel objected to the shackles. The objection was overruled. Ibid.; see also id., at 41–55. During the voir dire, Deck’s counsel renewed the objection. The objection was again overruled, the court stating that Deck “has been convicted and will remain in legirons and a belly chain.” Id., at 58. After the voir dire, Deck’s counsel once again objected, moving to strike the jury panel “because of the fact that Mr. Deck is shackled in front of the jury and makes them think that he is … violent today.” Id., at 58–59. The objection was again overruled, the court stating that his “being shackled takes any fear out of their minds.” Id., at 59. The penalty phase then proceeded with Deck in shackles. Deck was again sentenced to death. 136 S. W. 3d 481, 485 (Mo. 2004) (en banc). On appeal, Deck claimed that his shackling violated both Missouri law and the Federal Constitution. The Missouri Supreme Court rejected these claims, writing that there was “no record of the extent of the jury’s awareness of the restraints”; there was no “claim that the restraints impeded” Deck “from participating in the proceedings”; and there was “evidence” of “a risk” that Deck “might flee in that he was a repeat offender” who may have “killed his two victims to avoid being returned to custody.” Ibid. Thus, there was “sufficient evidence in the record to support the trial court’s exercise of its discretion” to require shackles, and in any event Deck “has not demonstrated that the outcome of his trial was prejudiced… . Neither being viewed in shackles by the venire panel prior to trial, nor being viewed while restrained throughout the entire trial, alone, is proof of prejudice.” Ibid. The court rejected Deck’s other claims of error and affirmed the sentence. We granted certiorari to review Deck’s claim that his shackling violated the Federal Constitution. II We first consider whether, as a general matter, the Constitution permits a State to use visible shackles routinely in the guilt phase of a criminal trial. The answer is clear: The law has long forbidden routine use of visible shackles during the guilt phase; it permits a State to shackle a criminal defendant only in the presence of a special need. This rule has deep roots in the common law. In the 18th century, Blackstone wrote that “it is laid down in our antient books, that, though under an indictment of the highest nature,” a defendant “must be brought to the bar without irons, or any manner of shackles or bonds; unless there be evident danger of an escape.” 4 W. Blackstone, Commentaries on the Laws of England 317 (1769) (footnote omitted); see also 3 E. Coke, Institutes of the Laws of England *34 (“If felons come in judgement to answer, … they shall be out of irons, and all manner of bonds, so that their pain shall not take away any manner of reason, nor them constrain to answer, but at their free will”). Blackstone and other English authorities recognized that the rule did not apply at “the time of arraignment,” or like proceedings before the judge. Blackstone, supra, at 317; see also Trial of Christopher Layer, 16 How. St. Tr. 94, 99 (K. B. 1722). It was meant to protect defendants appearing at trial before a jury. See King v. Waite, 1 Leach 28, 36, 168 Eng. Rep. 117, 120 (K. B. 1743) (“[B]eing put upon his trial, the Court immediately ordered [the defendant’s] fetters to be knocked off”). American courts have traditionally followed Blackstone’s “ancient” English rule, while making clear that “in extreme and exceptional cases, where the safe custody of the prisoner and the peace of the tribunal imperatively demand, the manacles may be retained.” 1 J. Bishop, New Criminal Procedure §955, p. 573 (4th ed. 1895); see also id., at 572–573 (“[O]ne at the trial should have the unrestrained use of his reason, and all advantages, to clear his innocence… . Our American courts adhere pretty closely to this doctrine” (internal quotation marks omitted)); State v. Roberts, 86 N. J. Super. 159, 163–165, 206 A. 2d 200, 203 (1965); French v. State, 377 P. 2d 501, 502–504 (Okla. Crim. App. 1962); Eaddy v. People, 115 Colo. 488, 490, 174 P. 2d 717, 718 (1946) (en banc); State v. McKay, 63 Nev. 118, 153–158, 165 P. 2d 389, 405–406 (1946); Blaine v. United States, 136 F. 2d 284, 285 (CADC 1943) (per curiam); Blair v. Commonwealth, 171 Ky. 319, 327–329, 188 S. W. 390, 393 (App. 1916); Hauser v. People, 210 Ill. 258, 264–267, 71 N. E. 416, 421 (1904); Parker v. Territory, 5 Ariz. 283, 287, 52 P. 361, 363 (1898); State v. Williams, 18 Wash. 47, 48–50, 50 P. 580, 581 (1897); Rainey v. State, 20 Tex. Ct. App. 455, 472–473 (1886) (opinion of White, P. J.); State v. Smith, 11 Ore. 205, 8 P. 343, 343 (1883); Poe v. State, 78 Tenn. 673, 674–678 (1882); State v. Kring, 64 Mo. 591, 592 (1877); People v. Harrington, 42 Cal. 165, 167 (1871); see also F. Wharton, Criminal Pleading and Practice §540a, p. 369 (8th ed. 1880); 12 Cyclopedia of Law and Procedure 529 (1904). While these earlier courts disagreed about the degree of discretion to be afforded trial judges, see post, at 9–14 (Thomas, J., dissenting), they settled virtually without exception on a basic rule embodying notions of fundamental fairness: trial courts may not shackle defendants routinely, but only if there is a particular reason to do so. More recently, this Court has suggested that a version of this rule forms part of the Fifth and Fourteenth Amendments’ due process guarantee. Thirty-five years ago, when considering the trial of an unusually obstreperous criminal defendant, the Court held that the Constitution sometimes permitted special measures, including physical restraints. Allen, 397 U. S., at 343–344. The Court wrote that “binding and gagging might possibly be the fairest and most reasonable way to handle” such a defendant. Id., at 344. But the Court immediately added that “even to contemplate such a technique … arouses a feeling that no person should be tried while shackled and gagged except as a last resort.” Ibid. Sixteen years later, the Court considered a special courtroom security arrangement that involved having uniformed security personnel sit in the first row of the courtroom’s spectator section. The Court held that the Constitution allowed the arrangement, stating that the deployment of security personnel during trial is not “the sort of inherently prejudicial practice that, like shackling, should be permitted only where justified by an essential state interest specific to each trial.” Holbrook, 475 U. S., at 568–569. See also Estelle v. Williams, 425 U. S. 501, 503, 505 (1976) (making a defendant appear in prison garb poses such a threat to the “fairness of the factfinding process” that it must be justified by an “essential state policy”). Lower courts have treated these statements as setting forth a constitutional standard that embodies Blackstone’s rule. Courts and commentators share close to a consensus that, during the guilt phase of a trial, a criminal defendant has a right to remain free of physical restraints that are visible to the jury; that the right has a constitutional dimension; but that the right may be overcome in a particular instance by essential state interests such as physical security, escape prevention, or courtroom decorum. See, e.g., Dyas v. Poole, 309 F. 3d 586, 588–589 (CA9 2002) (per curiam); Harrell v. Israel, 672 F. 2d 632, 635 (CA7 1982) (per curiam); State v. Herrick, 324 Mont. 76, 78–82, 101 P. 3d 755, 757–759 (2004); Hill v. Commonwealth, 125 S. W. 3d 221, 233–234 (Ky. 2004); State v. Turner, 143 Wash. 2d 715, 723–727, 23 P. 3d 499, 504–505 (2001) (en banc); Myers v. State, 2000 OK CR 25, ¶19, 17 P. 3d 1021, 1033; State v. Shoen, 598 N. W. 2d 370, 374–377 (Minn. 1999); Lovell v. State, 347 Md. 623, 635–645, 702 A. 2d 261, 268–272 (1997); People v. Jackson, 14 Cal. App. 4th 1818, 1822–1830, 18 Cal. Rptr. 2d 586, 588–594 (1993); Cooks v. State, 844 S. W. 2d 697, 722 (Tex. Crim. App. 1992); State v. Tweedy, 219 Conn. 489, 504–508, 594 A. 2d 906, 914–915 (1991); State v. Crawford, 99 Idaho 87, 93–98, 577 P. 2d 1135, 1141–1146 (1978); People v. Brown, 45 Ill. App. 3d 24, 26–28, 358 N. E. 2d 1362, 1363–1364 (1977); State v. Tolley, 290 N. C. 349, 362–371, 226 S. E. 2d 353, 365–369 (1976); see also 21A Am. Jur. 2d, Criminal Law §§1016, 1019 (1998); see generally J. Krauskopf, Physical Restraint of the Defendant in the Courtroom, 15 St. Louis U. L. J. 351 (1970–1971); ABA Standards for Criminal Justice: Discovery and Trial by Jury 15–3.2, pp. 188–191 (3d ed. 1996). Lower courts have disagreed about the specific procedural steps a trial court must take prior to shackling, about the amount and type of evidence needed to justify restraints, and about what forms of prejudice might warrant a new trial, but they have not questioned the basic principle. They have emphasized the importance of preserving trial court discretion (reversing only in cases of clear abuse), but they have applied the limits on that discretion described in Holbrook, Allen, and the early English cases. In light of this precedent, and of a lower court consensus disapproving routine shackling dating back to the 19th century, it is clear that this Court’s prior statements gave voice to a principle deeply embedded in the law. We now conclude that those statements identify a basic element of the “due process of law” protected by the Federal Constitution. Thus, the Fifth and Fourteenth Amendments prohibit the use of physical restraints visible to the jury absent a trial court determination, in the exercise of its discretion, that they are justified by a state interest specific to a particular trial. Such a determination may of course take into account the factors that courts have traditionally relied on in gauging potential security problems and the risk of escape at trial. III We here consider shackling not during the guilt phase of an ordinary criminal trial, but during the punishment phase of a capital case. And we must decide whether that change of circumstance makes a constitutional difference. To do so, we examine the reasons that motivate the guilt-phase constitutional rule and determine whether they apply with similar force in this context. A Judicial hostility to shackling may once primarily have reflected concern for the suffering—the “tortures” and “torments”—that “very painful” chains could cause. Krauskopf, supra, at 351, 353 (internal quotation marks omitted); see also Riggins v. Nevada, 504 U. S. 127, 154, n. 4 (1992) (Thomas, J., dissenting) (citing English cases curbing the use of restraints). More recently, this Court’s opinions have not stressed the need to prevent physical suffering (for not all modern physical restraints are painful). Instead they have emphasized the importance of giving effect to three fundamental legal principles. First, the criminal process presumes that the defendant is innocent until proved guilty. Coffin v. United States, 156 U. S. 432, 453 (1895) (presumption of innocence “lies at the foundation of the administration of our criminal law”). Visible shackling undermines the presumption of innocence and the related fairness of the factfinding process. Cf. Estelle, supra, at 503. It suggests to the jury that the justice system itself sees a “need to separate a defendant from the community at large.” Holbrook, supra, at 569; cf. State v. Roberts, 86 N. J. Super., at 162, 206 A. 2d, at 202 (“[A] defendant ‘ought not be brought to the Bar in a contumelious Manner; as with his Hands tied together, or any other Mark of Ignominy and Reproach … unless there be some Danger of a Rescous [rescue] or Escape’ ” (quoting 2 W. Hawkins, Pleas of the Crown, ch. 28, §1, p. 308 (1716–1721) (section on arraignments))). Second, the Constitution, in order to help the accused secure a meaningful defense, provides him with a right to counsel. See, e.g., Amdt. 6; Gideon v. Wainwright, 372 U. S. 335, 340–341 (1963). The use of physical restraints diminishes that right. Shackles can interfere with the accused’s “ability to communicate” with his lawyer. Allen, 397 U. S., at 344. Indeed, they can interfere with a defendant’s ability to participate in his own defense, say by freely choosing whether to take the witness stand on his own behalf. Cf. Cranburne’s Case, 13 How. St. Tr. 222 (K. B. 1696) (“Look you, keeper, you should take off the prisoners irons when they are at the bar, for they should stand at their ease when they are tried” (footnote omitted)); People v. Harrington, 42 Cal., at 168 (shackles “impos[e] physical burdens, pains, and restraints … , … ten[d] to confuse and embarrass” defendants’ “mental faculties,” and thereby tend “materially to abridge and prejudicially affect his constitutional rights”). Third, judges must seek to maintain a judicial process that is a dignified process. The courtroom’s formal dignity, which includes the respectful treatment of defendants, reflects the importance of the matter at issue, guilt or innocence, and the gravity with which Americans consider any deprivation of an individual’s liberty through criminal punishment. And it reflects a seriousness of purpose that helps to explain the judicial system’s power to inspire the confidence and to affect the behavior of a general public whose demands for justice our courts seek to serve. The routine use of shackles in the presence of juries would undermine these symbolic yet concrete objectives. As this Court has said, the use of shackles at trial “affront[s]” the “dignity and decorum of judicial proceedings that the judge is seeking to uphold.” Allen, supra, at 344; see also Trial of Christopher Layer, 16 How. St. Tr., at 99 (statement of Mr. Hungerford) (“[T]o have a man plead for his life” in shackles before “a court of justice, the highest in the kingdom for criminal matters, where the king himself is supposed to be personally present” undermines the “dignity of the Court”). There will be cases, of course, where these perils of shackling are unavoidable. See Allen, supra, at 344. We do not underestimate the need to restrain dangerous defendants to prevent courtroom attacks, or the need to give trial courts latitude in making individualized security determinations. We are mindful of the tragedy that can result if judges are not able to protect themselves and their courtrooms. But given their prejudicial effect, due process does not permit the use of visible restraints if the trial court has not taken account of the circumstances of the particular case. B The considerations that militate against the routine use of visible shackles during the guilt phase of a criminal trial apply with like force to penalty proceedings in capital cases. This is obviously so in respect to the latter two considerations mentioned, securing a meaningful defense and maintaining dignified proceedings. It is less obviously so in respect to the first consideration mentioned, for the defendant’s conviction means that the presumption of innocence no longer applies. Hence shackles do not undermine the jury’s effort to apply that presumption. Nonetheless, shackles at the penalty phase threaten related concerns. Although the jury is no longer deciding between guilt and innocence, it is deciding between life and death. That decision, given the “ ‘severity’ ” and “ ‘finality’ ” of the sanction, is no less important than the decision about guilt. Monge v. California, 524 U. S. 721, 732 (1998) (quoting Gardner v. Florida, 430 U. S. 349, 357 (1977)). Neither is accuracy in making that decision any less critical. The Court has stressed the “acute need” for reliable decisionmaking when the death penalty is at issue. Monge, supra, at 732 (citing Lockett v. Ohio, 438 U. S. 586, 604 (1978) (plurality opinion)). The appearance of the offender during the penalty phase in shackles, however, almost inevitably implies to a jury, as a matter of common sense, that court authorities consider the offender a danger to the community—often a statutory aggravator and nearly always a relevant factor in jury decisionmaking, even where the State does not specifically argue the point. Cf. Brief for Respondent 25–27. It also almost inevitably affects adversely the jury’s perception of the character of the defendant. See Zant v. Stephens, 462 U. S. 862, 900 (1983) (Rehnquist, J., concurring in judgment) (character and propensities of the defendant are part of a “unique, individualized judgment regarding the punishment that a particular person deserves”). And it thereby inevitably undermines the jury’s ability to weigh accurately all relevant considerations—considerations that are often unquantifiable and elusive—when it determines whether a defendant deserves death. In these ways, the use of shackles can be a “thumb [on] death’s side of the scale.” Sochor v. Florida, 504 U. S. 527, 532 (1992) (internal quotation marks omitted); see also Riggins, 504 U. S., at 142 (Kennedy, J., concurring) (through control of a defendant’s appearance, the State can exert a “powerful influence on the outcome of the trial”). Given the presence of similarly weighty considerations, we must conclude that courts cannot routinely place defendants in shackles or other physical restraints visible to the jury during the penalty phase of a capital proceeding. The constitutional requirement, however, is not absolute. It permits a judge, in the exercise of his or her discretion, to take account of special circumstances, including security concerns, that may call for shackling. In so doing, it accommodates the important need to protect the courtroom and its occupants. But any such determination must be case specific; that is to say, it should reflect particular concerns, say special security needs or escape risks, related to the defendant on trial. IV Missouri claims that the decision of its high court meets the Constitution’s requirements in this case. It argues that the Missouri Supreme Court properly found: (1) that the record lacks evidence that the jury saw the restraints; (2) that the trial court acted within its discretion; and, in any event, (3) that the defendant suffered no prejudice. We find these arguments unconvincing. The first argument is inconsistent with the record in this case, which makes clear that the jury was aware of the shackles. See App. 58–59 (Deck’s attorney stated on the record that “Mr. Deck [was] shackled in front of the jury”) (emphasis added); id., at 59 (trial court responded that “him being shackled takes any fear out of their minds”). The argument also overstates the Missouri Supreme Court’s holding. The court said, “[T]rial counsel made no record of the extent of the jury’s awareness of the restraints throughout the penalty phase, and Appellant does not claim that the restraints impeded him from participating in the proceedings.” 136 S. W. 3d, at 485 (emphasis added). This statement does not suggest that the jury was unaware of the restraints. Rather, it refers to the degree of the jury’s awareness, and hence to the kinds of prejudice that might have occurred. The second argument—that the trial court acted within its discretion—founders on the record’s failure to indicate that the trial judge saw the matter as one calling for discretion. The record contains no formal or informal findings. Cf. supra, at 9 (requiring a case-by-case determination). The judge did not refer to a risk of escape—a risk the State has raised in this Court, see Tr. of Oral Arg. 36–37—or a threat to courtroom security. Rather, he gave as his reason for imposing the shackles the fact that Deck already “has been convicted.” App. 58. While he also said that the shackles would “take any fear out of” the juror’s “minds,” he nowhere explained any special reason for fear. Id., at 59. Nor did he explain why, if shackles were necessary, he chose not to provide for shackles that the jury could not see—apparently the arrangement used at trial. If there is an exceptional case where the record itself makes clear that there are indisputably good reasons for shackling, it is not this one. The third argument fails to take account of this Court’s statement in Holbrook that shackling is “inherently prejudicial.” 475 U. S., at 568. That statement is rooted in our belief that the practice will often have negative effects, but—like “the consequences of compelling a defendant to wear prison clothing” or of forcing him to stand trial while medicated—those effects “cannot be shown from a trial transcript.” Riggins, supra, at 137. Thus, where a court, without adequate justification, orders the defendant to wear shackles that will be seen by the jury, the defendant need not demonstrate actual prejudice to make out a due process violation. The State must prove “beyond a reasonable doubt that the [shackling] error complained of did not contribute to the verdict obtained.” Chapman v. California, 386 U. S. 18, 24 (1967). V For these reasons, the judgment of the Missouri Supreme Court is reversed, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered.
THOMAS, J., DISSENTING
DECK V. MISSOURI
544 U. S. ____ (2005)
SUPREME COURT OF THE UNITED STATES
CARMAN L. DECK, PETITIONER v. MISSOURI on writ of certiorari to the supreme court of missouri [May 23, 2005] Justice Thomas, with whom Justice Scalia joins, dissenting. Carman Deck was convicted of murdering and robbing an elderly couple. He stood before the sentencing jury not as an innocent man, but as a convicted double murderer and robber. Today this Court holds that Deck’s due process rights were violated when he appeared at sentencing in leg irons, handcuffs, and a belly chain. The Court holds that such restraints may only be used where the use is “ ‘justified by an essential state interest’ ” that is “specific to the defendant on trial,” ante, at 1, and that is supported by specific findings by the trial court. Tradition—either at English common law or among the States—does not support this conclusion. To reach its result, the Court resurrects an old rule the basis for which no longer exists. It then needlessly extends the rule from trials to sentencing. In doing so, the Court pays only superficial heed to the practice of States and gives conclusive force to errant dicta sprinkled in a trio of this Court’s cases. The Court’s holding defies common sense and all but ignores the serious security issues facing our courts. I therefore respectfully dissent. I Carman Deck and his sister went to the home of Zelma and James Long on a summer evening in 1996. After waiting for nightfall, Deck and his sister knocked on the door of the Longs’ home, and when Mrs. Long answered, they asked for directions. Mrs. Long invited them in, and she and Mr. Long assisted them with directions. When Deck moved toward the door to leave, he drew a pistol, pointed it at the Longs, and ordered them to lie face down on their bed. The Longs did so, offering up money and valuables throughout the house and all the while begging that he not harm them. After Deck finished robbing their house, he stood at the edge of their bed, deliberating for 10 minutes over whether to spare them. He ignored their pleas and shot them each twice in the head. Deck later told police that he shot the Longs because he thought that they would be able to recognize him. Deck was convicted of the murders and robbery of the Longs and sentenced to death. The death sentence was overturned on appeal. Deck then had another sentencing hearing, at which he appeared in leg irons, a belly chain, and handcuffs. At the hearing, the jury heard evidence of Deck’s numerous burglary and theft convictions and his assistance in a jailbreak by two prisoners. On resentencing, the jury unanimously found six aggravating factors: Deck committed the murders while engaged in the commission of another unlawful homicide; Deck murdered each victim for the purpose of pecuniary gain; each murder involved depravity of mind; each murder was committed for the purpose of avoiding a lawful arrest; each murder was committed while Deck was engaged in a burglary; and each murder was committed while Deck was engaged in a robbery. The jury recommended, and the trial court imposed, two death sentences. Deck sought postconviction relief from his sentence, asserting among other things, that his due process and equal protection rights were violated by the trial court’s requirement that he appear in shackles. The Missouri Supreme Court rejected that claim. 136 S. W. 3d 481 (2004) (en banc). The court reasoned that “there was a risk that [Deck] might flee in that he was a repeat offender and evidence from the guilt phase of his trial indicated that he killed his two victims to avoid being returned to custody,” and thus it could not conclude that the trial court had abused its discretion. Id., at 485. II My legal obligation is not to determine the wisdom or the desirability of shackling defendants, but to decide a purely legal question: Does the Due Process Clause of the Fourteenth Amendment preclude the visible shackling of a defendant. Therefore, I examine whether there is a deeply rooted legal principle that bars that practice. Medina v. California, 505 U. S. 437, 446 (1992); Apprendi v. New Jersey, 530 U. S. 466, 500 (2000) (Thomas, J., concurring); see also Chicago v. Morales, 527 U. S. 41, 102–106 (1999) (Thomas, J., dissenting). As I explain below, although the English common law had a rule against trying a defendant in irons, the basis for the rule makes clear that it should not be extended by rote to modern restraints, which are dissimilar in certain essential respects to the irons that gave rise to the rule. Despite the existence of a rule at common law, state courts did not even begin to address the use of physical restraints until the 1870’s, and the vast majority of state courts would not take up this issue until the 20th century, well after the ratification of the Fourteenth Amendment. Neither the earliest case nor the more modern cases reflect a consensus that would inform our understanding of the requirements of due process. I therefore find this evidence inconclusive. A English common law in the 17th and 18th centuries recognized a rule against bringing the defendant in irons to the bar for trial. See, e.g., 4 W. Blackstone, Commentaries on the Laws of England 317 (1769); 3 Coke, Institutes of the Laws of England *34 (1797) (hereinafter Coke). This rule stemmed from none of the concerns to which the Court points, ante, at 7–9—the presumption of innocence, the right to counsel, concerns about decorum, or accuracy in decisionmaking. Instead, the rule ensured that a defendant was not so distracted by physical pain during his trial that he could not defend himself. As one source states, the rule prevented prisoners from “any Torture while they ma[de] their defence, be their Crime never so great.” J. Kelyng, A Report of Divers Cases in Pleas of the Crown 10 (1708).[Footnote 1] This concern was understandable, for the irons of that period were heavy and painful. In fact, leather strips often lined the irons to prevent them from rubbing away a defendant’s skin. T. Gross, Manacles of the World: A Collector’s Guide to International Handcuffs, Leg Irons and other Miscellaneous Shackles and Restraints 25 (1997). Despite Coke’s admonition that “[i]t [was] an abuse that prisoners be chained with irons, or put to any pain before they be attained,” Coke *34, suspected criminals often wore irons during pretrial confinement. J. Langbein, The Origins of Adversary Criminal Trial 50, and n. 197 (2003) (hereinafter Langbein). For example, prior to his trial in 1722 for treason, Christopher Layer spent his confinement in irons. Layer’s counsel urged that his irons be struck off, for they allowed him to “sleep but in one posture.” Trial of Christopher Layer, 16 How. St. Tr. 94, 98 (K. B. 1722). The concern that felony defendants not be in severe pain at trial was acute because, before the 1730’s, defendants were not permitted to have the assistance of counsel at trial, with an early exception made for those charged with treason. Langbein 170–172. Instead, the trial was an “ ‘accused speaks’ ” trial, at which the accused defended himself. The accused was compelled to respond to the witnesses, making him the primary source of information at trial. Id., at 48; see also Faretta v. California, 422 U. S. 806, 823–824 (1975). As the Court acknowledges, ante, at 3, the rule against shackling did not extend to arraignment.[Footnote 2] A defendant remained in irons at arraignment because “he [was] only called upon to plead by advice of his counsel;” he was not on trial, where he would play the main role in defending himself. Trial of Christopher Layer, supra, at 100 (emphasis added). A modern-day defendant does not spend his pretrial confinement wearing restraints. The belly chain and handcuffs are of modest, if not insignificant, weight. Neither they nor the leg irons cause pain or suffering, let alone pain or suffering that would interfere with a defendant’s ability to assist in his defense at trial. And they need not interfere with a defendant’s ability to assist his counsel—a defendant remains free to talk with counsel during trial, and restraints can be employed so as to ensure that a defendant can write to his counsel during the trial. Restraints can also easily be removed when a defendant testifies, so that any concerns about testifying can be ameliorated. Modern restraints are therefore unlike those that gave rise to the traditional rule. The Court concedes that modern restraints are nothing like the restraints of long ago, ante, at 7, and even that the rule at common law did not rest on any of the “three fundamental legal principles” the Court posits to support its new rule, ibid. Yet the Court treats old and modern restraints as similar for constitutional purposes merely because they are both types of physical restraints. This logical leap ignores that modern restraints do not violate the principle animating the common-law rule. In making this leap, the Court’s strays from the appropriate legal inquiry of examining common-law traditions to inform our understanding of the Due Process Clause. B In the absence of a common-law rule that applies to modern-day restraints, state practice is also relevant to determining whether a deeply rooted tradition supports the conclusion that the Fourteenth Amendment’s Due Process Clause limits shackling. See Morales, 527 U. S., at 102–106 (Thomas, J., dissenting). The practice among the States, however, does not support, let alone require, the conclusion that shackling can be done only where “particular concerns … related to the defendant on trial” are articulated as findings in the record. Ante, at 11. First, state practice is of modern, not longstanding, vintage. The vast majority of States did not address the issue of physical restraints on defendants during trial until the 20th century. Second, the state cases—both the earliest to address shackling and even the later cases—reflect substantial differences that undermine the contention that the Due Process Clause so limits the use of physical restraints. Third, state- and lower federal-court cases decided after Illinois v. Allen, 397 U. S. 337 (1970), Estelle v. Williams, 425 U. S. 501 (1976), and Holbrook v. Flynn, 475 U. S. 560 (1986), are not evidence of a current consensus about the use of physical restraints. Such cases are but a reflection of the dicta contained in Allen, Estelle, and Holbrook. 1 State practice against shackling defendants was established in the 20th century. In 35 States, no recorded state-court decision on the issue appears until the 20th century.[Footnote 3] Of those 35 States, 21 States have no recorded decision on the question until the 1950’s or later.[Footnote 4] The 14 state (including then-territorial) courts that addressed the matter before the 20th century only began to do so in the 1870’s.[Footnote 5] The California Supreme Court’s decision in People v. Harrington, 42 Cal. 165 (1871), “seems to have been the first case in this country where this ancient rule of the common law was considered and enforced.” State v. Smith, 11 Ore. 205, 208, 8 P. 343 (1883). The practice in the United States is thus of contemporary vintage. State practice that was only nascent in the late 19th century is not evidence of a consistent unbroken tradition dating to the common law, as the Court suggests. Ante, at 3–4. The Court does not even attempt to account for the century of virtual silence between the practice established at English common law and the emergence of the rule in the United States. Moreover, the belated and varied state practice is insufficient to warrant the conclusion that shackling of a defendant violates his due process rights. See Martinez v. Court of Appeal of Cal., Fourth Appellate Dist., 528 U. S. 152, 159 (2000) (where no history of a right to appeal much before the 20th century, no historical support for a right to self-representation on appeal). 2 The earliest state cases reveal courts’ divergent views of visible shackling, undermining the notion that due process cabins shackling to cases in which “particular concerns … related to the defendant on trial” are supported by findings on the record. Ante, at 11. The Supreme Court of the New Mexico Territory held that great deference was to be accorded the trial court’s decision to put the defendant in shackles, permitting a reviewing court to presume that there had been a basis for doing so if the record lay silent. Territory of New Mexico v. Kelly, 2 N. M. 292, 304–306 (1882). Only if the record “affirmatively” showed “no reason whatever” for shackling was the decision to shackle a defendant erroneous. Ibid.; see State v. Allen, 45 W. Va. 65, ___, 30 S. E. 209, 211 (1898) (following Kelly), overruled in relevant part, State v. Brewster, 164 W. Va. 173, 182, 261 S. E. 2d 77, 82 (1979). The Alabama Supreme Court also left the issue to the trial court’s discretion and went so far as to bar any appeal from the trial court’s decision to restrain the defendant. Faire v. State, 58 Ala. 74, 80–81 (1877); see Poe v. State, 78 Tenn. 673, 677 (1882) (decision to manacle a defendant during trial “left to the sound discretion of the trial court” and subject to abuse-of-discretion standard of review). Mississippi concluded that the decision to shackle a defendant “may be safely committed to courts and sheriffs, whose acts are alike open to review in the courts and at the ballot box.”[Footnote 6] Lee v. State, 51 Miss. 566, ___ (1875), overruled on other grounds Wingo v. State, 62 Miss. 311 (1884). By contrast, California, Missouri, Washington, and Oregon adopted more restrictive approaches. In People v. Harrington, supra, the California Supreme Court held that shackling a defendant “without evident necessity” of any kind violated the common-law rule as well as state law and was prejudicial to the defendant. Id., at 168–169. A few years later, the Missouri courts took an even more restrictive view, concluding that the use of shackles or other such restraints was permitted only if warranted by the defendant’s conduct “at the time of trial.” State v. Kring, 64 Mo. 591, 593 (1877); see State v. Smith, supra, at 207–208, 8 P., at 343 (following Kring and Harrington without discussion); State v. Williams, 18 Wash. 47, 50–51, 50 P. 580, 581–582 (1897) (adopting Kring’s test). Texas took an intermediate position. The Texas Court of Appeals relied on Kring, and at the same time deferred to the decision made by the sheriff to bring the defendant into the courtroom in shackles. See Rainey v. State, 20 Tex. Ct. App. 455, 472 (1886); see also Parker v. Territory, 5 Ariz. 283, 287–288, 52 P. 361, 363 (1898) (following Harrington but permitting the shackling of a defendant at arraignment based on the crime for which he had been arrested as well as the reward that had been offered for his recapture). Thus, in the late 19th century States agreed that generally defendants ought to come to trial unfettered, but they disagreed over the breadth of discretion to be afforded trial courts. A bare majority of States required that trial courts and even jailers be given great leeway in determining when a defendant should be restrained; a minority of States severely constrained such discretion, in some instances by limiting the information that could be considered; and an even smaller set of States took an intermediate position. While the most restrictive view adopted by States is perhaps consistent with the rule Deck seeks, the majority view is flatly inconsistent with requiring a State to show and for a trial court to set forth findings of an “ ‘essential state interest’ ” “specific to the defendant on trial” before shackling a defendant. Ante, at 1. In short, there was no consensus that supports elevating the rule against shackling to a federal constitutional command. 3 The modern cases provide no more warrant for the Court’s approach than do the earliest cases. The practice in the 20th century did not resolve the divisions among States that emerged in the 19th century. As more States addressed the issue, they continued to express a general preference that defendants be brought to trial without shackles. They continued, however, to disagree about the latitude to be given trial courts. Many deferred to the judgment of the trial court,[Footnote 7] and some to the views of those responsible for guarding the defendant.[Footnote 8] States also continued to disagree over whether the use of shackles was inherently prejudicial.[Footnote 9] Moreover, States differed over the information that could be considered in deciding to shackle the defendant and the certainty of the risk that had to be established, with a small minority limiting the use of shackles to instances arising from conduct specific to the particular trial or otherwise requiring an imminent threat.[Footnote 10] The remaining States permitted courts to consider a range of information outside the trial, including past escape,[Footnote 11] prior convictions,[Footnote 12] the nature of the crime for which the defendant was on trial,[Footnote 13] conduct prior to trial while in prison,[Footnote 14] any prior disposition toward violence,[Footnote 15] and physical attributes of the defendant, such as his size, physical strength, and age.[Footnote 16] The majority permits courts to continue to rely on these factors, which are undeniably probative of the need for shackling, as a basis for shackling a defendant both at trial and at sentencing. Ante, at 6–7. In accepting these traditional factors, the Court rejects what has been adopted by few States—that courts may consider only a defendant’s conduct at the trial itself or other information demonstrating that it is a relative certainty that the defendant will engage in disruptive or threatening conduct at his trial. See State v. Coursolle, 255 Minn. 384, 389, 97 N. W. 2d 472, 477 (1959) (defining “immediate necessity” to be demonstrated only by the defendant’s conduct “at the time of the trial”); State v. Finch, 137 Wash. 2d 792, 850, 975 P. 2d 967, 1001 (1999) (en banc); Blair v. Commonwealth, 171 Ky. 319, 327–328, 188 S. W. 390, 393 (1916); State v. Temple, 194 Mo. 237, ___, 92 S. W. 869, 872 (1906); but see 136 S. W. 3d, at 485 (case below) (appearing to have abandoned this test). A number of those traditional factors were present in this case. Here, Deck killed two people to avoid arrest, a fact to which he had confessed. Evidence was presented that Deck had aided prisoners in an escape attempt. Moreover, a jury had found Deck guilty of two murders, the facts of which not only make this crime heinous but also demonstrate a propensity for violence. On this record, and with facts found by a jury, the Court says that it needs more. Since the Court embraces reliance on the traditional factors supporting the use of visible restraints, its only basis for reversing is the requirement of specific on-the-record findings by the trial judge. This requirement is, however, inconsistent with the traditional discretion afforded to trial courts and is unsupported by state practice. This additional requirement of on-the-record findings about that which is obvious from the record makes little sense to me. 4 In recent years, more of a consensus regarding the use of shackling has developed, with many courts concluding that shackling is inherently prejudicial. But rather than being firmly grounded in deeply rooted principles, that consensus stems from a series of ill-considered dicta in Illinois v. Allen, 397 U. S. 337 (1970), Estelle v. Williams, 425 U. S. 501 (1976), and Holbrook v. Flynn, 475 U. S. 560 (1986). In Allen, the trial court had removed the defendant from the courtroom until the court felt he could conform his conduct to basic standards befitting a court proceeding. 397 U. S., at 340–341. This Court held that removing the defendant did not violate his due process right to be present for his trial. In dicta, the Court suggested alternatives to removal, such as citing the defendant for contempt or binding and gagging him. Id., at 344. The Court, however, did express some revulsion at the notion of binding and gagging a defendant. Ibid. Estelle and Holbrook, repeated Allen’s dicta. Estelle, supra, at 505; Holbrook, supra, at 568. The Court in Holbrook went one step further than it had in Allen, describing shackling as well as binding and gagging in dicta as “inherently prejudicial.” 475 U. S., at 568. The current consensus that the Court describes is one of its own making. Ante, at 5. It depends almost exclusively on the dicta in this Court’s opinions in Holbrook, Estelle, and Allen. Every lower court opinion the Court cites as evidence of this consensus traces its reasoning back to one or more of these decisions.[Footnote 17] These lower courts were interpreting this Court’s dicta, not reaching their own independent consensus about the content of the Due Process Clause. More important, these decisions represent recent practice, which does not determine whether the Fourteenth Amendment, as properly and traditionally interpreted, i.e., as a statement of law, not policy preferences, embodies a right to be free from visible, painless physical restraints at trial. III Wholly apart from the propriety of shackling a defendant at trial, due process does not require that a defendant remain free from visible restraints at the penalty phase of a capital trial. Such a requirement has no basis in tradition or even modern state practice. Treating shackling at sentencing as inherently prejudicial ignores the commonsense distinction between a defendant who stands accused and a defendant who stands convicted. A There is no tradition barring the use of shackles or other restraints at sentencing. Even many modern courts have concluded that the rule against visible shackling does not apply to sentencing. See, e.g., State v. Young, 853 P. 2d 327, 350 (Utah 1993); Duckett v. State, 104 Nev. 6, 11, 752 P. 2d 752, 755 (1988) (per curiam); State v. Franklin, 97 Ohio St. 3d 1, 18–19, 776 N. E. 2d 26, 46–47 (2002); but see Bello v. State, 547 So. 2d 914, 918 (Fla. 1989) (applying rule against shackling at sentencing, but suggesting that “lesser showing of necessity” may be appropriate). These courts have rejected the suggestion that due process imposes such limits because they have understood the difference between a man accused and a man convicted. See, e.g., Young, supra, at 350; Duckett, supra, at 11, 752 P. 2d, at 755. This same understanding is reflected even in the guilt-innocence phase. In instances in which the jury knows that the defendant is an inmate, though not yet convicted of the crime for which he is on trial, courts have frequently held that the defendant’s status as inmate ameliorates any prejudice that might have flowed from the jury seeing him in handcuffs.[Footnote 18] The Court’s decision shuns such common sense. B In the absence of a consensus with regard to the use of visible physical restraints even in modern practice, we should not forsake common sense in determining what due process requires. Capital sentencing jurors know that the defendant has been convicted of a dangerous crime. It strains credulity to think that they are surprised at the sight of restraints. Here, the jury had already concluded that there was a need to separate Deck from the community at large by convicting him of double murder and robbery. Deck’s jury was surely aware that Deck was jailed; jurors know that convicted capital murderers are not left to roam the streets. It blinks reality to think that seeing a convicted capital murderer in shackles in the courtroom could import any prejudice beyond that inevitable knowledge. Jurors no doubt also understand that it makes sense for a capital defendant to be restrained at sentencing. By sentencing, a defendant’s situation is at its most dire. He no longer may prove himself innocent, and he faces either life without liberty or death. Confronted with this reality, a defendant no longer has much to lose—should he attempt escape and fail, it is still lengthy imprisonment or death that awaits him. For any person in these circumstances, the reasons to attempt escape are at their apex. A defendant’s best opportunity to do so is in the courtroom, for he is otherwise in jail or restraints. See Westman, Handling the Problem Criminal Defendant in the Courtroom: The Use of Physical Restraints and Expulsion in the Modern Era, 2 San Diego Justice J. 507, 526–527 (1994) (hereinafter Westman). In addition, having been convicted, a defendant may be angry. He could turn that ire on his own counsel, who has failed in defending his innocence. See, e.g., State v. Forrest, 609 S. E. 2d 241, 248–249 (N. C. App. 2005) (defendant brutally attacked his counsel at sentencing). Or, for that matter, he could turn on a witness testifying at his hearing or the court reporter. See, e.g., People v. Byrnes, 33 N. Y. 2d 343, 350, 308 N. E. 2d 435, 438 (1974) (defendant lunged at witness during trial); State v. Harkness, 252 Kan. 510, 516, 847 P. 2d 1191, 1197 (1993) (defendant attacked court reporter at arraignment). Such thoughts could well enter the mind of any defendant in these circumstances, from the most dangerous to the most docile. That a defendant now convicted of his crimes appears before the jury in shackles thus would be unremarkable to the jury. To presume that such a defendant suffers prejudice by appearing in handcuffs at sentencing does not comport with reality. IV The modern rationales proffered by the Court for its newly minted rule likewise fail to warrant the conclusion that due process precludes shackling at sentencing. Moreover, though the Court purports to be mindful of the tragedy that can take place in a courtroom, the stringent rule it adopts leaves no real room for ensuring the safety of the courtroom. A Although the Court offers the presumption of innocence as a rationale for the modern rule against shackling at trial, it concedes the presumption has no application at sentencing. Ante, at 9. The Court is forced to turn to the far more amorphous need for “accuracy” in sentencing. Ibid. It is true that this Court’s cases demand reliability in the factfinding that precedes the imposition of a sentence of death. Monge v. California, 524 U. S. 721, 732 (1998). But shackles may undermine the factfinding process only if seeing a convicted murderer in them is prejudicial. As I have explained, this farfetched conjecture defies the reality of sentencing. The Court baldly asserts that visible physical restraints could interfere with a defendant’s ability to participate in his defense. Ante, at 7–8. I certainly agree that shackles would be impermissible if they were to seriously impair a defendant’s ability to assist in his defense, Riggins v. Nevada, 504 U. S. 127, 154, n. 4 (1992) (Thomas, J., dissenting), but there is no evidence that shackles do so. Deck does not argue that the shackles caused him pain or impaired his mental faculties. Nor does he argue that the shackles prevented him from communicating with his counsel during trial. Counsel sat next to him; he remained fully capable of speaking with counsel. Likewise, Deck does not claim that he was unable to write down any information he wished to convey to counsel during the course of the trial. Had the shackles impaired him in that way, Deck could have sought to have at least one of his hands free to make it easier for him to write. Courts have permitted such arrangements. See, e.g., People v. Alvarez, 14 Cal. 4th 155, 191, 926 P. 2d 365, 386 (1996); State v. Jimerson, 820 S. W. 2d 500, 502 (Mo. App. 1991). The Court further expresses concern that physical restraints might keep a defendant from taking the stand on his own behalf in seeking the jury’s mercy. Ante, at 7–8. But this concern is, again, entirely hypothetical. Deck makes no claim that, but for the physical restraints, he would have taken the witness stand to plead for his life. And under the rule the Court adopts, Deck and others like him need make no such assertion, for prejudice is presumed absent a showing by the government to the contrary. Even assuming this concern is real rather than imagined, it could be ameliorated by removing the restraints if the defendant wishes to take the stand. See, e.g., De Wolf v. State, 96 Okla. Cr. 382, 383, 256 P. 2d 191, 193 (1953) (leg irons removed from defendant in capital case when he took the witness stand). Instead, the Court says, the concern requires a categorical rule that the use of visible physical restraints violates the Due Process Clause absent a demanding showing. The Court’s solution is overinclusive. The Court also asserts the rule it adopts is necessary to protect courtroom decorum, which the use of shackles would offend. Ante, at 8. This courtroom decorum rationale misunderstands this Court’s precedent. No decision of this Court has ever intimated, let alone held, that the protection of the “courtroom’s formal dignity,” ibid., is an individual right enforceable by criminal defendants. Certainly, courts have always had the inherent power to ensure that both those who appear before them and those who observe their proceedings conduct themselves appropriately. See, e.g., Estes v. Texas, 381 U. S. 532, 540–541 (1965). The power of the courts to maintain order, however, is not a right personal to the defendant, much less one of constitutional proportions. Far from viewing the need for decorum as a right the defendant can invoke, this Court has relied on it to limit the conduct of defendants, even when their constitutional rights are implicated. This is why a defendant who proves himself incapable of abiding by the most basic rules of the court is not entitled to defend himself, Faretta v. California, 422 U. S., at 834–835, n. 46, or to remain in the courtroom, see Allen, 397 U. S., at 343. The concern for courtroom decorum is not a concern about defendants, let alone their right to due process. It is a concern about society’s need for courts to operate effectively. Wholly apart from the unwarranted status the Court accords “courtroom decorum,” the Court fails to explain the affront to the dignity of the courts that the sight of physical restraints poses. I cannot understand the indignity in having a convicted double murderer and robber appear before the court in visible physical restraints. Our Nation’s judges and juries are exposed to accounts of heinous acts daily, like the brutal murders Deck committed in this case. Even outside the courtroom, prisoners walk through courthouse halls wearing visible restraints. Courthouses are thus places in which members of the judiciary and the public come into frequent contact with defendants in restraints. Yet, the Court says, the appearance of a convicted criminal in a belly chain and handcuffs at a sentencing hearing offends the sensibilities of our courts. The courts of this Nation do not have such delicate constitutions. Finally, the Court claims that “[t]he appearance of the offender during the penalty phase in shackles … almost inevitably implies to a jury, as a matter of common sense, that court authorities consider the offender a danger to the community—often a statutory aggravator and nearly always a relevant factor in jury decisionmaking.” Ante, at 10. This argument is flawed. It ignores the fact that only relatively recently have the penalty and guilt phases been conducted separately. That the historical evidence reveals no consensus prohibiting visible modern-day shackles during capital trials suggests that there is similarly no consensus prohibiting shackling during capital sentencing. Moreover, concerns about a defendant’s dangerousness exist at the guilt phase just as they exist at the penalty phase—jurors will surely be more likely to convict a seemingly violent defendant of murder than a seemingly placid one. If neither common law nor modern state cases support the Court’s position with respect to the guilt phase, I see no reason why the fact that a defendant may be perceived as a future danger would support the Court’s position with respect to the penalty phase. B The Court expresses concern for courtroom security, but its concern rings hollow in light of the rule it adopts. The need for security is real. Judges face the possibility that a defendant or his confederates might smuggle a weapon into court and harm those present, or attack with his bare hands. For example, in 1999, in Berks County, Pennsylvania, a “defendant forced his way to the bench and beat the judge unconscious.” Calhoun, Violence Toward Judicial Officials, 576 Annals of the American Academy of Political and Social Science 54, 61 (2001). One study of Pennsylvania judges projected that over a 20-year career, district justices had a 31 percent probability of being physically assaulted one or more times. See Harris, Kirschner, Rozek, & Weiner, Violence in the Judicial Workplace: One State’s Experience, 576 Annals of the American Academy of Political and Social Science 38, 42 (2001). Judges are not the only ones who face the risk of violence. Sheriffs and courtroom bailiffs face the second highest rate of homicide in the workplace, a rate which is 15 times higher than the national average. Faust & Raffo, Local Trial Court Response to Courthouse Safety, 576 Annals of the American Academy of Political and Social Science 91, 93–94 (2001); Weiner et al., Safe and Secure: Protecting Judicial Officials, 36 Court Review 26, 27 (Winter 2000). The problem of security may only be worsening. According to the General Accounting Office (GAO), the nature of the prisoners in the federal system has changed: “[T]here are more ‘hard-core tough guys’ and more multiple-defendant cases,” making the work of the federal marshals increasingly difficult. GAO, Federal Judiciary Security: Comprehensive Risk-Based Program Should Be Fully Implemented 21 (July 1994). Security issues are particularly acute in state systems, in which limited manpower and resources often leave judges to act as their own security. See Harris, supra, at 46. Those resources further vary between rural and urban areas, with many rural areas able to supply only minimal security. Security may even be at its weakest in the courtroom itself, for there the defendant is the least restrained. Westman 526. In the face of this real danger to courtroom officials and bystanders, the Court limits the use of visible physical restraints to circumstances “specific to a particular trial,” ante, at 6, i.e., “particular concerns … related to the defendant on trial,” ante, at 11. Confining the analysis to trial-specific circumstances precludes consideration of limits on the security resources of courts. Under that test, the particulars of a given courthouse (being nonspecific to any particular defendant) are irrelevant, even if the judge himself is the only security, or if a courthouse has few on-duty officers standing guard at any given time, or multiple exits. Forbidding courts from considering such circumstances fails to accommodate the unfortunately dire security situation faced by this Nation’s courts. *** The Court’s decision risks the lives of courtroom personnel, with little corresponding benefit to defendants. This is a risk that due process does not require. I respectfully dissent.