delivered the opinion of the Court.

The question presented in this case is what a defendant must show in order to demonstrate a Sixth Amendment violation where the trial court fails to inquire into a potential conflict of interest about which it knew or reasonably should have known.

I

In 1993, a Virginia jury convicted petitioner Mickens of the premeditated murder of Timothy Hall during or following the commission of an attempted forcible sodomy. Finding the murder outrageously and wantonly vile, it sentenced petitioner to death. In June 1998, Mickens filed a petition for writ of habeas corpus, see 28 U. S. C. § 2254 (1994 ed. and Supp. V), in the United States District Court for the Eastern District of Virginia, alleging, inter alia, that he was denied effective assistance of counsel because one of his court-appointed attorneys had a conflict of interest at trial. Federal habeas counsel had discovered that petitioner’s lead trial attorney, Bryan Saunders, was representing Hall (the victim) on assault and concealed-weapons charges at the time of the murder. Saunders had been appointed to represent Hall, a juvenile, on March 20, 1992, and had met with him once for 15 to 30 minutes some time the following week. Hall’s body was discovered on March 30,1992, and four days *165later a juvenile court judge dismissed the charges against him, noting on the docket sheet that Hall was deceased. The one-page docket sheet also listed Saunders as Hall’s counsel. On April 6, 1992, the same judge appointed Saunders to represent petitioner. Saunders did not disclose to the court, his co-counsel, or petitioner that he had previously represented Hall. Under Virginia law, juvenile case files are confidential and may not generally be disclosed without a court order, see Va. Code Ann. § 16.1-305 (1999), but petitioner learned about Saunders’ prior representation when a clerk mistakenly produced Hall’s file to federal habeas counsel.

The District Court held an evidentiary hearing and denied petitioner’s habeas petition. A divided panel of the Court of Appeals for the Fourth Circuit reversed, 227 F. 3d 203 (2000), and the Court of Appeals granted rehearing en banc, 240 F. 3d 348 (2001). As an initial matter, the 7-to-3 en banc majority determined that petitioner’s failure to raise his conflict-of-interest claim in state court did not preclude review, concluding that petitioner had established cause and that the “inquiry as to prejudice for purposes of excusing [petitioner’s] default... incorporates the test for evaluating his underlying conflict of interest claim.” Id., at 356-357. On the merits, the Court of Appeals assumed that the juvenile court judge had neglected a duty to inquire into a potential conflict, but rejected petitioner’s argument that this failure either mandated automatic reversal of his conviction or relieved him of the burden of showing that a conflict of interest adversely affected his representation. Relying on Cuyler v. Sullivan, 446 U. S. 335 (1980), the court held that a defendant must show “both an actual conflict of interest and an adverse effect even if the trial court failed to inquire into a potential conflict about which it reasonably should have known,” 240 F. 3d, at 355-356. Concluding that petitioner had not demonstrated adverse effect, id., at 360, it affirmed the District Court’s denial of habeas relief. We *166granted a stay of execution of petitioner’s sentence and granted certiorari. 532 U. S. 970 (2001).

II

The Sixth Amendment provides that a criminal defendant shall have the right to “the Assistance of Counsel for his defence.” This right has been accorded, we have said, “not for its own sake, but because of the effect it has on the ability of the accused to receive a fair trial.” United States v. Cronic, 466 U. S. 648, 658 (1984). It follows from this that assistance which is ineffective in preserving fairness does not meet the constitutional mandate, see Strickland v. Washington, 466 U. S. 668, 685-686 (1984); and it also follows that defects in assistance that have no probable effect upon the trial’s outcome do not establish a constitutional violation. As a general matter, a defendant alleging a Sixth Amendment violation must demonstrate “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Id., at 694.

There is an exception to this general rule. We have spared the defendant the need of showing probable effect upon the outcome, and have simply presumed such effect, where assistance of counsel has been denied entirely or during a critical stage of the proceeding. When that has occurred, the likelihood that the verdict is unreliable is so high that a case-by-case inquiry is unnecessary. See Cronic, supra, at 658-659; see also Geders v. United States, 425 U. S. 80, 91 (1976); Gideon v. Wainwright, 372 U. S. 335, 344-345 (1963). But only in “circumstances of that magnitude” do we forgo individual inquiry into whether counsel’s inadequate performance undermined the reliability of the verdict. Cronic, supra, at 659, n. 26.

We have held in several cases that “circumstances of that magnitude” may also arise when the defendant’s attorney actively represented conflicting interests. The nub of the question before us is whether the principle established by these cases provides an exception to the general rule of *167Strickland under the circumstances of the present case. To answer that question, we must examine those cases in some detail.1

In Holloway v. Arkansas, 435 U. S. 475 (1978), defense counsel had objected that he could not adequately represent the divergent interests of three codefendants. Id., at 478-480. Without inquiry, the trial court had denied counsel’s motions for the appointment of separate counsel and had refused to allow counsel to cross-examine any of the defendants on behalf of the other two. The Holloway Court deferred to the judgment of counsel regarding the existence of a disabling conflict, recognizing that a defense attorney is in the best position to determine when a conflict exists, that he has an ethical obligation to advise the court of any problem, and that his declarations to the court are “virtually made *168under oath.” Id., at 485-486 (internal quotation marks omitted). Holloway presumed, moreover, that the conflict, “which [the defendant] and his counsel tried to avoid by timely objections to the joint representation,” id., at 490, undermined the adversarial process. The presumption was justified because joint representation of conflicting interests is inherently suspect, and because counsel’s conflicting obligations to multiple defendants “effectively sea[l] his lips on crucial matters” and make it difficult to measure the precise harm arising from counsel’s errors. Id., at 489-490. Holloway thus creates an automatic reversal rule only where defense counsel is forced to represent codefendants over his timely objection, unless the trial court has determined that there is no conflict. Id., at 488 (“[W]henever a trial court improperly requires joint representation over timely objection reversal is automatic”).

In Cuyler v. Sullivan, 446 U. S. 335 (1980), the respondent was one of three defendants accused of murder who were tried separately, represented by the same counsel. Neither counsel nor anyone else objected to the multiple representation, and counsel’s opening argument at Sullivan’s trial suggested that the interests of the defendants were aligned. Id., at 347-348. We declined to extend Holloway’s automatic reversal rule to this situation and held that, absent objection, a defendant must demonstrate that “a conflict of interest actually affected the adequacy of his representation.” 446 U. S., at 348-349. In addition to describing the defendant’s burden of proof, Sullivan addressed separately a trial court’s duty to inquire into the propriety of a multiple representation, construing Holloway to require inquiry only when “the trial court knows or reasonably should know that a particular conflict exists,” 446 U. S., at 3472 — which is not *169to be confused with when the trial court is aware of a vague, unspecified possibility of conflict, such as that which “inheres in almost every instance of multiple representation,” id., at 348. In Sullivan, no “special circumstances” triggered the trial court’s duty to inquire. Id., at 346.

Finally, in Wood v. Georgia, 450 U. S. 261 (1981), three indigent defendants convicted of distributing obscene materials had their probation revoked for failure to make the requisite $500 monthly payments on their $5,000 fines. We granted certiorari to consider whether this violated the Equal Protection Clause, but during the course of our consideration certain disturbing circumstances came to our attention: At the probation-revocation hearing (as at all times since their arrest) the defendants had been represented by the lawyer for their employer (the owner of the business that purveyed the obscenity), and their employer paid the attorney’s fees. The employer had promised his employees he would pay their fines, and had generally kept that promise but had not done so in these defendants’ case. This record suggested that the employer’s interest in establishing a favorable equal-protection precedent (reducing the fines he would have to pay for his indigent employees in the future) diverged from the defendants’ interest in obtaining leniency or paying lesser fines to avoid imprisonment. Moreover, the possibility that counsel was actively representing the conflicting interests of employer and defendants “was sufficiently apparent at the time of the revocation hearing to impose upon the court a duty to inquire further.” Id., at 272. *170Because “[o]n the record before us, we [could not] be sure whether counsel was influenced in his basic strategic decisions by the interests of the employer who hired him,” ibid., we remanded for the trial court “to determine whether the conflict of interest that this record strongly suggests actually existed,” id., at 273.

Petitioner argues that the remand instruction in Wood established an “unambiguous rule” that where the trial judge neglects a duty to inquire into a potential conflict, the defendant, to obtain reversal of the judgment, need only show that his lawyer was subject to a conflict of interest, and need not show that the conflict adversely affected counsel’s performance. Brief for Petitioner 21.3 He relies upon the lan*171guage in the remand instruction directing the trial court to grant a new revocation hearing if it determines that “an actual conflict of interest existed,” Wood, supra, at 273, without requiring a further determination that the conflict adversely affected counsel’s performance. As used in the remand instruction, however, we think “an actual conflict of interest” meant precisely a conflict that affected counsel’s performance — as opposed to a mere theoretical division of loyalties. It was shorthand for the statement in Sullivan that “a defendant who shows that a conflict of interest actually affected the adequacy of his representation need not demonstrate prejudice in order to obtain relief.” 446 U. S., at 349-350 (emphasis added).4 This is the only interpreta*172tion consistent with the Wood Court’s earlier description of why it could not decide the case without a remand: “On the record before us, we cannot be sure whether counsel was influenced in his basic strategic decisions by the interests of the employer who hired him. If this was the case, the due process rights of petitioners were not respected . . . .” 450 U. S., at 272 (emphasis added). The notion that Wood created a new rule sub silentio — and in a case where certio-rari had been granted on an entirely different question, and the parties had neither briefed nor argued the conflict-of-interest issue — is implausible.5

Petitioner’s proposed rule of automatic reversal when there existed a conflict that did not affect counsel’s performance, but the trial judge failed to make the Sullivan-mandated inquiry, makes little policy sense. As discussed, the rule applied when the trial judge is not aware of the *173conflict (and thus not obligated to inquire) is that prejudice will be presumed only if the conflict has significantly affected counsel’s performance — thereby rendering the verdict unreliable, even though Strickland prejudice cannot be shown. See Sullivan, supra, at 348-349. The trial court’s awareness of a potential conflict neither renders it more likely that counsel’s performance was significantly affected nor in any other way renders the verdict unreliable. Cf. United States v. Cronic, 466 U. S., at 662, n. 31. Nor does the trial judge’s failure to make the Sullivan-mandated inquiry often make it harder for reviewing courts to determine conflict and effect, particularly since those courts may rely on evidence and testimony whose importance only becomes established at the trial.

Nor, finally, is automatic reversal simply an appropriate means of enforcing Sullivan’s mandate of inquiry. Despite Justice Souter’s belief that there must be a threat of sanction (to wit, the risk of conferring a windfall upon the defendant) in order to induce “resolutely obdurate” trial judges to follow the law, post, at 208, we do not presume that judges are as careless or as partial as those police officers who need the incentive of the exclusionary rule, see United States v. Leon, 468 U. S. 897, 916-917 (1984). And in any event, the Sullivan standard, which requires proof of effect upon representation but (once such effect is shown) presumes prejudice, already creates an “incentive” to inquire into a potential conflict. In those cases where the potential conflict is in fact an actual one, only inquiry will enable the judge to avoid all possibility of reversal by either seeking waiver or replacing a conflicted attorney. We doubt that the deterrence of “judicial dereliction” that would be achieved by an automatic reversal rule is significantly greater.

Since this was not a case in which (as in Holloway) counsel protested his inability simultaneously to represent multiple defendants; and since the trial court’s failure to make the Sullivan-mandated inquiry does not reduce the petitioner’s *174burden of proof; it was at least necessary, to void the conviction, for petitioner to establish that the conflict of interest adversely affected his counsel’s performance. The Court of Appeals having found no such effect, see 240 F. 3d, at 360, the denial of habeas relief must be affirmed.

Ill

Lest today’s holding be misconstrued, we note that the only question presented was the effect of a trial court’s failure to inquire into a potential conflict upon the Sullivan rule that deficient performance of counsel must be shown. The case was presented and argued on the assumption that (absent some exception for failure to inquire) Sullivan would be applicable — requiring a showing of defective performance, but not requiring in addition (as Strickland does in other ineffectiveness-of-counsel cases), a showing of probable effect upon the outcome of trial. That assumption was not unreasonable in light of the holdings of Courts of Appeals, which have applied Sullivan “unblinkingly” to “all kinds of alleged attorney ethical conflicts,” Beets v. Scott, 65 F. 3d 1258, 1266 (CA5 1995) (en banc). They have invoked the Sullivan standard not only when (as here) there is a conflict rooted in counsel’s obligations to former clients, see, e. g., Perillo v. Johnson, 205 F. 3d 775, 797-799 (CA5 2000); Freund v. Butterworth, 165 F. 3d 839, 858-860 (CA11 1999); Mannhalt v. Reed, 847 F. 2d 576, 580 (CA9 1988); United States v. Young, 644 F. 2d 1008, 1013 (CA4 1981), but even when representation of the defendant somehow implicates counsel’s personal or financial interests, including a book deal, United States v. Hearst, 638 F. 2d 1190, 1193 (CA9 1980), a job with the prosecutor’s office, Garcia v. Bunnell, 33 F. 3d 1193, 1194-1195, 1198, n. 4 (CA9 1994), the teaching of classes to Internal Revenue Service agents, United States v. Michaud, 925 F. 2d 37, 40-42 (CA1 1991), a romantic “entanglement” with the prosecutor, Summerlin v. Stewart, 267 F. 3d 926, 935-941 (CA9 2001), or fear of antagonizing the *175trial judge, United States v. Sayan, 968 F. 2d 55, 64-65 (CADC 1992).

It must be said, however, that the language of Sullivan itself does not clearly establish, or indeed even support, such expansive application. “[Ujntil,” it said, “a defendant shows that his counsel actively represented conflicting interests, he has not established the constitutional predicate for his claim of ineffective assistance.” 446 U. S., at 350 (emphasis added). Both Sullivan itself, see id., at 348-349, and Holloway, see 435 U. S., at 490-491, stressed the high probability of prejudice arising from multiple concurrent representation, and the difficulty of proving that prejudice. See also Geer, Representation of Multiple Criminal Defendants: Conflicts of Interest and the Professional Responsibilities of the Defense Attorney, 62 Minn. L. Rev. 119,125-140 (1978); Lowenthal, Joint Representation in Criminal Cases: A Critical Appraisal, 64 Va. L. Rev. 939,941-950 (1978). Not all attorney conflicts present comparable difficulties. Thus, the Federal Rules of Criminal Procedure treat concurrent representation and prior representation differently, requiring a trial court to inquire into the likelihood of conflict whenever jointly charged defendants are represented by a single attorney (Rule 44(c)), but not when counsel previously represented another defendant in a substantially related matter, even where the trial court is aware of the prior representation.6 See Sullivan, supra, at 346, n. 10 (citing the Rule).

*176This is not to suggest that one ethical duty is more or less important than another. The purpose of our Holloway and Sullivan exceptions from the ordinary requirements of Strickland, however, is not to enforce the Canons of Legal Ethics, but to apply needed prophylaxis in situations where Strickland itself is evidently inadequate to assure vindication of the defendant’s Sixth Amendment right to counsel. See Nix v. Whiteside, 475 U. S. 157, 165 (1986) (“[B]reach of an ethical standard does not necessarily make out a denial of the Sixth Amendment guarantee of assistance of counsel”). In resolving this case on the grounds on which it was presented to us, we do not rule upon the need for the Sullivan prophylaxis in cases of successive representation. Whether Sullivan should be extended to such cases remains, as far as the jurisprudence of this Court is concerned, an open question.

* * *

For the reasons stated, the judgment of the Court of Appeals is

Affirmed.

concurring.

In its comprehensive analysis the Court has said all that is necessary to address the issues raised by the question presented, and I join the opinion in full. The trial judge’s failure to inquire into a suspected conflict is not the kind of error requiring a presumption of prejudice. We did not grant certiorari on a second question presented by petitioner: whether, if we rejected his proposed presumption, he had nonetheless established that a conflict of interest adversely affected his representation. I write separately to emphasize that the facts of this case well illustrate why a wooden rule requiring reversal is inappropriate for cases like this one.

*177At petitioner’s request, the District Court conducted an evidentiary hearing on the conflict claim and issued a thorough opinion, which found that counsel’s brief representation of the victim had no effect whatsoever on the course of petitioner’s trial. See Mickens v. Greene, 74 F. Supp. 2d 586 (ED Va. 1999). The District Court’s findings depend upon credibility judgments made after hearing the testimony of petitioner’s counsel, Bryan Saunders, and other witnesses. As a reviewing court, our role is not to speculate about counsel’s motives or about the plausibility of alternative litigation strategies. Our role is to defer to the District Court’s factual findings unless we can conclude they are clearly erroneous. See Lackawanna County District Attorney v. Coss, 532 U. S. 394, 406 (2001) (opinion of O’Connor, J.). The District Court found that Saunders did not believe he had any obligation to his former client, Timothy Hall, that would interfere with the litigation. See 74 F. Supp. 2d, at 606 (“[T]he Court concludes that, as a factual matter, Saunders did not believe that any continuing duties to a former client might interfere with his consideration of all facts and options for his current client” (internal quotation marks and alteration omitted)). Although the District Court concluded that Saunders probably did learn some matters that were confidential, it found that nothing the attorney learned was relevant to the subsequent murder case. See ibid. (“[T]he record here confirms that Saunders did not learn any confidential information from Hall that was relevant to Mickens’ defense either on the merits or at sentencing” (emphasis deleted)). Indeed, even if Saunders had learned relevant information, the District Court found that he labored under the impression he had no continuing duty at all to his deceased client. See id., at 605 (“[T]he record here reflects that, as far as Saunders was concerned, his allegiance to Hall, ‘[e]nded when I walked in the courtroom and they told me he was dead and the case was gone’”) (quoting Hearing Tr. 156-157, 218 (Jan. 13, 1999)). While Saunders’ belief *178may have been mistaken, it establishes that the prior representation did not influence the choices he made during the course of the trial. This conclusion is a good example of why a case-by-case inquiry is required, rather than simply adopting an automatic rule of reversal.

Petitioner’s description of roads not taken would entail two degrees of speculation. We would be required to assume that Saunders believed he had a continuing duty to the victim, and we then would be required to consider whether in this hypothetical case, the counsel would have been blocked from pursuing an alternative defense strategy. The District Court concluded that the prosecution’s case, coupled with the defendant's insistence on testifying, foreclosed the strategies suggested by petitioner after the fact. According to the District Court, there was no plausible argument that the victim consented to sexual relations with his murderer, given the bruises on the victim’s neck, blood marks showing the victim was stabbed before or during sexual intercourse, and, most important, petitioner’s insistence on testifying at trial that he had never met the victim. See 74 F. Supp. 2d, at 607 (“[T]he record shows that other facts foreclosed presentation of consent as a plausible alternative defense strategy”). The basic defense at the guilt phase was that petitioner was not at the scene; this is hardly consistent with the theory that there was a consensual encounter.

The District Court said the same for counsel’s alleged dereliction at the sentencing phase. Saunders’ failure to attack the character of the 17-year-old victim and his mother had nothing to do with the putative conflict of interest. This strategy was rejected as likely to backfire, not only by Saunders, but also by his co-counsel, who owed no duty to Hall. See id., at 608 (“[T]he record here dispels the contention that the failure to use negative information about Hall is attributable to any conflict of interest on the part of Saunders”). These facts, and others relied upon by the District Court, provide compelling evidence that a theoretical conflict does *179not establish a constitutional violation, even when the conflict is one about which the trial judge should have known.

The constitutional question must turn on whether trial counsel had a conflict of interest that hampered the representation, not on whether the trial judge should have been more assiduous in taking prophylactic measures. If it were otherwise, the judge’s duty would not be limited to cases where the attorney is suspected of harboring a conflict of interest. The Sixth Amendment protects the defendant against an ineffective attorney, as well as a conflicted one. See Strickland v. Washington, 466 U. S. 668, 685-686 (1984). It would be a major departure to say that the trial judge must step in every time defense counsel appears to be providing ineffective assistance, and indeed, there is no precedent to support this proposition. As the Sixth Amendment guarantees the defendant the assistance of counsel, the infringement of that right must depend on a deficiency of the lawyer, not of the trial judge. There is no reason to presume this guarantee unfulfilled when the purported conflict has had no effect on the representation.

With these observations, I join the opinion of the Court.

dissenting.

This case raises three uniquely important questions about a fundamental component of our criminal justice system— the constitutional right of a person accused of a capital offense to have the effective assistance of counsel for his defense.1 The first is whether a capital defendant’s attorney *180has a duty to disclose that he was representing the defendant’s alleged victim at the time of the murder. Second, is whether, assuming disclosure of the prior representation, the capital defendant has a right to refuse the appointment of the conflicted attorney. Third, is whether the trial judge, who knows or should know of such prior representation, has a duty to obtain the defendant’s consent before appointing that lawyer to represent him. Ultimately, the question presented by this case is whether, if these duties exist and if all of them are violated, there exist “circumstances that are so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjustified.” United States v. Cronic, 466 U. S. 648, 658 (1984).

I

The first critical stage in the defense of a capital case is the series of pretrial meetings between the accused and his counsel when they decide how the case should be defended. A lawyer cannot possibly determine how best to represent a new client unless that client is willing to provide the lawyer with a truthful account of the relevant facts. When an indigent defendant first meets his newly appointed counsel, he will often falsely maintain his complete innocence. Truthful disclosures of embarrassing or incriminating facts are contingent on the development of the client’s confidence in the undivided loyalty of the lawyer. Quite obviously, knowledge that the lawyer represented the victim would be a substantial obstacle to the development of such confidence.

It is equally true that a lawyer’s decision to conceal such an important fact from his new client would have comparable ramifications. The suppression of communication and truncated investigation that would unavoidably follow from such a decision would also make it difficult, if not altogether impossible, to establish the necessary level of trust that should *181characterize the “delicacy of relation” between attorney and client.2

In this very case, it is likely that Mickens misled his counsel, Bryan Saunders, given the fact that Mickens gave false testimony at his trial denying any involvement in the crime despite the overwhelming evidence that he had killed Timothy Hall after a sexual encounter. In retrospect, it seems obvious that the death penalty might have been avoided by acknowledging Mickens’ involvement, but emphasizing the evidence suggesting that their sexual encounter was consensual. Mickens’ habeas counsel garnered evidence suggesting that Hall was a male prostitute, App. 137,149, 162,169; that the area where Hall was killed was known for prostitution, id,., at 169-170; and that there was no evidence that Hall was forced to the secluded area where he was ultimately murdered. An unconflicted attorney could have put forward a defense tending to show that Mickens killed Hall only after the two engaged in consensual sex, but Saunders offered no such defense. This was a crucial omission — a finding of forcible sodomy was an absolute prerequisite to Mick-ens’ eligibility for the death penalty.3 Of course, since that *182strategy would have led to conviction of a noncapital offense, counsel would have been unable to persuade the defendant to divulge the information necessary to support such a defense and then ultimately to endorse the strategy unless he had earned the complete confidence of his client.

Saunders’ concealment of essential information about his prior representation of the victim was a severe lapse in his professional duty. The lawyer’s duty to disclose his representation of a client related to the instant charge is not only intuitively obvious, it is as old as the profession. Consider this straightforward comment made by Justice Story in 1824:

“An attorney is bound to disclose to his client every adverse retainer, and even every prior retainer, which may affect the discretion of the latter. No man can be supposed to be indifferent to the knowledge of facts, which work directly on his interests, or bear on the freedom of his choice of counsel. When a client employs an attorney, he has a right to presume, if the latter be silent on the point, that he has no engagements, which interfere, in any degree, with his exclusive devotion to the cause confided to him; that he has no interest, which may betray his judgment, or endanger his fidelity.” Williams v. Reed, 29 F. Cas. 1386,1390 (No. 17,733) (CC Me.).

Mickens’ lawyer’s violation of this fundamental obligation of disclosure is indefensible. The relevance of Saunders’ prior representation of Hall to the new appointment was far too important to be concealed.

II

If the defendant is found guilty of a capital offense, the ensuing proceedings that determine whether he will be put to death are critical in every sense of the word. At those proceedings, testimony about the impact of the crime on the victim, including testimony about the character of the victim, may have a critical effect on the jury’s decision. Payne v. *183Tennessee, 501 U. S. 808 (1991). Because a lawyer’s fiduciary relationship with his deceased client survives the client’s death, Swidler & Berlin v. United States, 524 U. S. 399 (1998), Saunders necessarily labored under conflicting obligations that were irreconcilable. He had a duty to protect the reputation and confidences of his deceased client, and a duty to impeach the impact evidence presented by the prosecutor.4

Saunders’ conflicting obligations to his deceased client, on the one hand, and to his living client, on the other, were unquestionably sufficient to give Mickens the right to insist on different representation.5 For the “right to counsel guaranteed by the Constitution contemplates the services of an attorney devoted solely to the interests of his client,” Von Moltke v. Gillies, 332 U. S. 708, 725 (1948).6 Moreover, in my judgment, the right to conflict-free counsel is just as firmly protected by the Constitution as the defendant’s right *184of self-representation recognized in Faretta v. California, 422 U. S. 806 (1975).7

Ill

When an indigent defendant is unable to retain his own lawyer, the trial judge’s appointment of counsel is itself a critical stage of a criminal trial. At that point in the proceeding, by definition, the defendant has no lawyer to protect his interests and must rely entirely on the judge. For that reason it is “the solemn duty of a . . . judge before whom a defendant appears without counsel to make a thorough inquiry and to take all steps necessary to insure the fullest protection of this constitutional right at every stage of the proceedings.” Von Moltke, 332 U. S., at 722.

This duty with respect to indigent defendants is far more imperative than the judge’s duty to investigate the possibility of a conflict that arises when retained counsel represents either multiple or successive defendants. It is true that in a situation of retained counsel, “[ujnless the trial court knows or reasonably should know that a particular conflict exists, the court need not initiate an inquiry.” Cuyler v. Sullivan, 446 U. S. 335, 347 (1980).8 But when, as was true in this *185case, the judge is not merely reviewing the permissibility of the defendants’ choice of counsel, but is responsible for making the choice herself, and when she knows or should know that a conflict does exist, the duty to make a thorough inquiry is manifest and unqualified.9 Indeed, under far less compelling circumstances, we squarely held that when a record discloses the “possibility of a conflict” between the interests of the defendants and the interests of the party paying their counsel’s fees, the Constitution imposes a duty of inquiry on the state-court judge even when no objection was made. Wood v. Georgia, 450 U. S. 261, 267, 272 (1981).

IV

Mickens had a constitutional right to the services of an attorney devoted solely to his interests. That right was violated. The lawyer who did represent him had a duty to disclose his prior representation of the victim to Mickens and to the trial judge. That duty was violated. When Mickens had no counsel, the trial judge had a duty to “make a thorough inquiry and to take all steps necessary to insure the fullest protection of” his right to counsel. Von Moltke, 332 *186U. S., at 722. Despite knowledge of the lawyer’s prior representation, she violated that duty.

We will never know whether Mickens would have received the death penalty if those violations had not occurred nor precisely what effect they had on Saunders’ representation of Mickens.10 We do know that he did not receive the kind of representation that the Constitution guarantees. If Mickens had been represented by an attorney-impostor who never passed a bar examination, we might also be unable to determine whether the impostor’s educational shortcomings “‘actually affected the adequacy of his representation.’” Ante, at 171 (emphasis deleted). We would, however, surely set aside his conviction if the person who had represented him was not a real lawyer. Four compelling reasons make setting aside the conviction the proper remedy in this case.

First, it is the remedy dictated by our holdings in Holloway v. Arkansas, 435 U. S. 475 (1978), Cuyler v. Sullivan, 446 U. S. 335 (1980), and Wood v. Georgia, 450 U. S. 261 (1981). In this line of precedent, our focus was properly upon the duty of the trial court judge to inquire into a potential conflict. This duty was triggered either via defense counsel’s objection, as was the case in Holloway, or some other “special circumstances” whereby the serious potential for conflict was brought to the attention of the trial court judge. Sullivan, 446 U. S., at 346. As we unambiguously stated in Wood, “Sullivan mandates a reversal when the trial court has failed to make an inquiry even though it ‘knows or reasonably should know that a particular conflict *187exists.’ ” 450 U. S., at 272, n. 18. It is thus wrong for the Court to interpret Justice Powell’s language as referring only to a division of loyalties “that affected counsel’s performance.” Ante, at 171, and n. 3 (emphasis deleted).11 Wood nowhere hints of this meaning of “actual conflict of interest” 450 U. S., at 273, nor does it reference Sullivan in “shorthand,” ante, at 171. Rather, Wood cites Sullivan explicitly in order to make a factual distinction: In a circumstance, such as in Wood, in which the judge knows or should know of the conflict, no showing of adverse effect is required. But when, as in Sullivan, the judge lacked this knowledge, such a showing is required. Wood, 450 U. S., at 272-274.12

*188Second, it is the only remedy that responds to the real possibility that Mickens would not have received the death penalty if he had been represented by conflict-free counsel during the critical stage of the proceeding in which he first met with his lawyer. We should presume that the lawyer for the victim of a brutal homicide is incapable of establishing the kind of relationship with the defendant that is essential to effective representation.

Third, it is the only remedy that is consistent with the legal profession’s historic and universal condemnation of the representation of conflicting interests without the full disclosure and consent of all interested parties.13 The Court’s novel and naive assumption that a lawyer’s divided loyalties *189are acceptable unless it can be proved that they actually affected counsel’s performance is demeaning to the profession.

Finally, “justice must satisfy the appearance of justice.” Offutt v. United States, 348 U. S. 11, 14 (1954). Setting aside Mickens’ conviction is the only remedy that can maintain public confidence in the fairness of the procedures employed in capital cases. Death is a different kind of punishment from any other that may be imposed in this country. “From the point of view of the defendant, it is different in both its severity and its finality. From the point of view of society, the action of the sovereign in taking the life of one of its citizens also differs dramatically from any other legitimate state action. It is of vital importance to the defendant and to the community 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, 357-358 (1977). A rule that allows the State to foist a murder victim’s lawyer onto his accused is not only capricious; it poisons the integrity of our adversary system of justice.

I respectfully dissent.

dissenting.

A judge who knows or should know that counsel for a criminal defendant facing, or engaged in, trial has a potential conflict of interests is obliged to enquire into the potential conflict and assess its threat to the fairness of the proceeding. See Wheat v. United States, 486 U. S. 153, 160 (1988); Wood v. Georgia, 450 U. S. 261, 272 (1981); Cuyler v. Sullivan, 446 U. S. 335, 347 (1980). Cf. Holloway v. Arkansas, 435 U. S. 475, 484 (1978). Unless the judge finds that the risk of inadequate representation is too remote for further concern, or finds that the defendant has intelligently assumed the risk and waived any potential Sixth or Fourteenth Amendment claim of inadequate counsel, the court must see that the lawyer is replaced. See ibid.; Glasser v. United States, 315 U. S. 60, 70 (1942). Cf. Wheat, supra, at 162; Ad*190visory Committee’s Notes on 1979 Amendments to Fed. Rule Crim. Proc. 44(c), 18 U. S. C. App., p. 1655.

The District Judge reviewing the federal habeas petition in this case found that the state judge who appointed Bryan Saunders to represent petitioner Mickens on a capital murder charge knew or should have known that obligations stemming from Saunders’s prior representation of the victim, Timothy Hall, potentially conflicted with duties entailed by defending Mickens.1 Mickens v. Greene, 74 F. Supp. 2d 586, 613-615 (ED Va. 1999). The state judge was therefore obliged to look further into the extent of the risk and, if necessary, either secure Mickens’s knowing and intelligent assumption of the risk or appoint a different lawyer. The state judge, however, did nothing to discharge her constitutional duty of care. Id., at 614. In the one case in which we have devised a remedy for such judicial dereliction, we held that the ensuing judgment of conviction must be reversed and the defendant afforded a new trial. Holloway, *191supra, at 491; see also Wood, supra, at 272, n. 18. That should be the result here.

I

The Court today holds, instead, that Mickens should be denied this remedy because Saunders failed to employ a formal objection as a means of bringing home to the appointing judge the risk of conflict. Ante, at 173-174. Without an objection, the majority holds, Mickens should get no relief absent a showing that the risk turned into an actual conflict with adverse effect on the representation provided to Mick-ens at trial. Ibid. But why should an objection matter when even without an objection the state judge knew or should have known of the risk and was therefore obliged to enquire further? What would an objection have added to the obligation the state judge failed to honor? The majority says that in circumstances like those now before us, we have already held such an objection necessary for reversal, absent proof of actual conflict with adverse effect, so that this case calls simply for the application of precedent, albeit precedent not very clearly stated. Ante, at 171-172.

The majority’s position is error, resting on a mistaken reading of our cases. Three are on point, Holloway v. Arkansas, supra; Cuyler v. Sullivan, supra; and Wood v. Georgia, supra.

In Holloway, a trial judge appointed one public defender to represent three criminal defendants tried jointly. 435 U. S., at 477. Three weeks before trial, counsel moved for separate representation; the court held a hearing and denied the motion. Ibid. The lawyer moved again for appointment of separate counsel before the jury was empaneled, on the ground that one or two of the defendants were considering testifying at trial, in which event the one lawyer’s ability to cross-examine would be inhibited. Id., at 478. The court again denied his motion. Ibid. After the prosecution rested, counsel objected to the joint representation a third time, advising the court that all three defendants had de*192cided to testify; again the court refused to appoint separate lawyers. Id., at 478-480. The defendants gave inconsistent testimony and were convicted on all counts. Id., at 481.

This Court held that the motions apprised the trial judge of a “risk” that continuing the joint representation would subject defense counsel in the pending trial to the impossible obligations of simultaneously furthering the conflicting interests of the several defendants, id., at 484, and we reversed the convictions on the basis of the judge’s failure to respond to the prospective conflict, without any further showing of harm, id., at 491. In particular, we rejected the argument that a defendant tried subject to such a disclosed risk should have to show actual prejudice caused by subsequent conflict. Id., at 488. We pointed out that conflicts created by multiple representation characteristically deterred a lawyer from taking some step that he would have taken if unconflicted, and we explained that the consequent absence of footprints would often render proof of prejudice virtually impossible. Id., at 489-491.

Next came Sullivan, involving multiple representation by two retained lawyers of three defendants jointly indicted but separately tried, 446 U. S., at 337. Sullivan, the defendant at the first trial, had consented to joint representation by the same lawyers retained by the two other accused, because he could not afford counsel of his own. Ibid. Sullivan was convicted of murder; the other two were acquitted in their subsequent trials. Id., at 338. Counsel made no objection to the multiple representation before or during trial, ibid.; nor did the convicted defendant argue that the trial judge otherwise knew or should have known of the risk described in Holloway, that counsel’s representation might be impaired by conflicting obligations to the defendants to be tried later, 446 U. S., at 343.

This Court held that multiple representation did not raise enough risk of impaired representation in a coming trial to *193trigger a trial court’s duty to enquire further, in the absence of “special circumstances.”2 Id., at 346. The most obvious special circumstance would be an objection. See Holloway, supra, at 488. Indeed, because multiple representation was not suspect per se, and because counsel was in the best position to anticipate a risk of conflict, the Court spoke at one point as though nothing but an objection would place a court on notice of a prospective conflict. Sullivan, 446 U. S., at 348 (“[A] defendant who raised no objection at trial must demonstrate that an actual conflict of interest adversely affected his lawyer’s performance” (footnote omitted)). But the Court also explained that courts must rely on counsel in “large measure,” id., at 347, that is, not exclusively, and it spoke in general terms of a duty to enquire that arises when “the trial court knows or reasonably should know that a particular conflict exists,”3 ibid, (footnote omitted). Accord*194ingly, the Court did not rest the result simply on the failure of counsel to object, but said instead that “[n]othing in the circumstances of this case indicates that the trial court had a duty to inquire whether there was a conflict of interest,” ibid. For that reason, it held respondent bound to show “that a conflict of interest actually affected the adequacy of his representation.” Id., at 349.

The different burdens on the Holloway and Sullivan defendants are consistent features of a coherent scheme for dealing with the problem of conflicted defense counsel; a prospective risk of conflict subject to judicial notice is treated differently from a retrospective claim that a completed proceeding was tainted by conflict, although the trial judge had not been derelict in any duty to guard against it. When the problem comes to the trial court’s attention before any potential conflict has become actual, the court has a duty to act prospectively to assess the risk and, if the risk is not too remote, to eliminate it or to render it acceptable through a defendant’s knowing and intelligent waiver. This duty is something more than the general responsibility to rule without committing legal error; it is an affirmative obligation to investigate a disclosed possibility that defense counsel will be unable to act with uncompromised loyalty to his client. It was the judge’s failure to fulfill that duty of care to enquire further and do what might be necessary that the Holloway Court remedied by vacating the defendant’s subsequent conviction. 435 U. S., at 487, 491. The error occurred when the judge failed to act, and the remedy restored the defend*195ant to the position he would have occupied if the judge had taken reasonable steps to fulfill his obligation. But when the problem of conflict comes to judicial attention not prospectively, but only after the fact, the defendant must show an actual conflict with adverse consequence to him in order to get relief. Sullivan, supra, at 349. Fairness requires nothing more, for no judge was at fault in allowing a trial to proceed even though fraught with hidden risk.

In light of what the majority holds today, it bears repeating that, in this coherent scheme established by Holloway and Sullivan, there is nothing legally crucial about an objection by defense counsel to tell a trial judge that conflicting interests may impair the adequacy of counsel’s representation. Counsel’s objection in Holloway was important as a fact sufficient to put the judge on notice that he should en-quire. In most multiple-representation cases, it will take just such an objection to alert a trial judge to prospective conflict, and the Sullivan Court reaffirmed that the judge is obliged to take reasonable prospective action whenever a timely objection is made. 446 U. S., at 346. But the Court also indicated that an objection is not required as a matter of law: “Unless the trial court knows or reasonably should know that a particular conflict exists, the court need not initiate an enquiry.” Id., at 347. The Court made this clear beyond cavil 10 months later when Justice Powell, the same Justice who wrote the Sullivan opinion, explained in Wood v. Georgia that Sullivan “mandates a reversal when the trial court has failed to make an inquiry even though it ‘knows or reasonably should know that a particular conflict exists.’” 450 U. S., at 272, n. 18 (emphasis in original).

Since the District Court in this case found that the state judge was on notice of a prospective potential conflict, 74 F. Supp. 2d, at 613-615, this case calls for nothing more than the application of the prospective notice rule announced and exemplified by Holloway and confirmed in Sullivan and Wood. The remedy for the judge’s dereliction of duty *196should be an order vacating the conviction and affording a new trial.

But in the majority’s eyes, this conclusion takes insufficient account of Wood, whatever may have been the sensible scheme staked out by Holloway and Sullivan, with a defendant’s burden turning on whether a court was apprised of a conflicts problem prospectively or retrospectively. The majority says that Wood holds that the distinction is between cases where counsel objected and all other cases, regardless of whether a trial court was put on notice prospectively in some way other than by an objection on the record. See ante, at 172-174. In Wood, according to the majority, the trial court had notice, there was no objection on the record, arid the defendant was required to show actual conflict and adverse effect.

Wood is not easy to read, and I believe the majority misreads it. The first step toward seeing where the majority goes wrong is to recall that the Court in Wood said outright what I quoted before, that Sullivan “mandates a reversal when the trial court has failed to make an inquiry even though it ‘knows or reasonably should know that a particular conflict exists.’ ” 450 U. S., at 272, n. 18. This statement of a trial judge’s obligation, like the statement in Sullivan that it quoted, 446 U. S., at 347, said nothing about the need for an objection on the record. True, says the majority, but the statement was dictum to be disregarded as “inconsistent” with Wood’s holding. Ante, at 168-169, n. 2. This is a polite way of saying that the Wood Court did not know what it was doing; that it stated the general rule of reversal for failure to enquire when on notice (as in Holloway), but then turned around and held that such a failure called for reversal only when the defendant demonstrated an actual conflict (as in Sullivan).

This is not what happened. Wood did not hold that in the absence of objection, the Sullivan rule governs even when a judge is prospectively on notice of a risk of conflicted counsel. *197Careful attention to Wood shows that the case did not involve prospective notice of risk unrealized, and that it held nothing about the general rule to govern in such circumstances. What Wood did decide was how to deal with a possible conflict of interests that becomes known to the trial court only at the conclusion of the trial proceeding at which it may have occurred, and becomes known not to a later habeas court but to the judge who handed down sentences at trial, set probation 19 months later after appeals were exhausted, and held a probation revocation proceeding 4 months after that.4

The Wood defendants were convicted of distributing obscene material as employees of an adult bookstore and theater, after trials at which they were defended by privately retained counsel. 450 U. S., at 262-263. They were each ordered to pay fines and sentenced to 12-month prison terms that were suspended in favor of probation on the condition that they pay their fines in installments, which they failed to do. Id., at 263-264. The Wood Court indicated that by the end of the proceeding to determine whether probation should be revoked because of the defendants’ failure to pay, the judge was on notice that defense counsel might have been laboring under a conflict between the interests of the defendant employees and those of their employer, possibly as early as the time the sentences were originally handed down nearly two years earlier, App. 11-16 in Wood v. Georgia, O. T. 1979, No. 79-6027 (Mar. 18, 1977, sentencing). See Wood, 450 U. S., at 272 (“at the revocation hearing, or at earlier stages of the proceedings below”). The fines were so high that the original sentencing assumption must have been that the store and theater owner would pay them; defense counsel was paid by the employer, at least during the trial; the State *198pointed out a possible conflict to the judge;5 and counsel was attacking the fines with an equal protection argument, which weakened the strategy more obviously in the defendants’ interest, of requesting the court to reduce the fines or defer their collection. Id., at 272-273. This was enough, according to the Wood Court, to tell the judge that defense counsel may have been acting to further the- owner’s desire for a test case on equal protection, rather than the defendants’ interests in avoiding ruinous fines or incarceration. Ibid.

What is significant is that, as this Court thus described the circumstances putting the judge on notice, they were not complete until the revocation hearing was finished (nearly *199two years after sentencing) and the judge knew that the lawyer was relying heavily on equal protection instead of arguments for leniency to help the defendants. The Court noted that counsel stated he had sent a letter to the trial court after sentencing, saying the fines were more than the defendants could afford, id,., at 268, n. 13, a move obviously in the defendants’ interest. On the other hand, a reference to “equal protection,” which the Court could have taken as a reflection of the employer’s interest, did not occur until the very end of the revocation hearing. See App. A to Brief in Opposition in Wood v. Georgia, O. T. 1979, No. 79-6027, at 72 (transcript of Jan. 26, 1979, probation revocation hearing).6 The Wood Court also knew that a motion stressing equal protection was not filed by defense counsel until two weeks after the revocation hearing, on the day before probation was to be revoked and the defendants locked up, App. 35-36 in Wood v. Georgia, O. T. 1979, No. 79-6027 (Joint Motions to Modify Conditions of Probation Order — Filed Feb. 12,1979). 450 U. S., at 268. Since, in the Court’s view, counsel’s emphasis on the equal protection claim was one of the facts that *200together put the judge on notice of something amiss, and since the record shows that it was not clear that counsel was favoring the equal protection argument until, at the earliest, the very close of the revocation hearing, and more likely the day he filed his motion two weeks later, the Court could only have meant that the judge was put on notice of a conflict that may actually have occurred, not of a potential conflict that might occur later.7 At that point, as the Court saw it, there were only two further facts the judge would have needed to know to determine whether there had been an actual disqualifying conflict, and those were whether a concern for the interest of the employer had weakened the lawyer’s arguments for leniency, and whether the defendants had been informed of the conflict and waived their rights to uncon-flicted counsel.

This Court, of course, was in no position to resolve these remaining issues in the first instance. Whether the lawyer’s failure to press more aggressively for leniency was caused by a conflicting interest, for example, had never been explored at the trial level and there was no record to consult on the point.8 In deciding what to do, the Wood Court had *201two established procedural models to look to: Holloway’s procedure of vacating judgment9 when a judge had failed to enquire into a prospective conflict, and Sullivan’s procedure of determining whether the conflict that may well have occurred had actually occurred with some adverse effect.

Treating the case as more like Sullivan and remanding was obviously the correct choice. Wood was not like Holloway, in which the judge was put on notice of a risk before trial, that is, a prospective possibility of conflict. It was, rather, much closer to Sullivan, since any notice to a court went only to a conflict, if there was one, that had pervaded a completed trial proceeding extending over two years. The only difference between Wood and Sullivan was that, in Wood, the signs that a conflict may have occurred were clear to the judge at the close of the probation revocation proceeding, whereas the claim of conflict in Sullivan was not raised until after judgment in a separate habeas proceeding, see 446 U. S., at 338. The duty of the Wood judge could only have been to enquire into the past (what had happened two years earlier at sentencing, the setting of probation 19 months later, the ensuing failures to pay, and the testimony that had already been given at the revocation hearing), just like the responsibility of the state and federal habeas courts reviewing the record in Sullivan in postconviction proceedings, see 446 U. S., at 338-339. Since the Wood judge’s duty was unlike the Holloway judge’s obligation to take care for the future, it would have made no sense for the Wood Court to impose a Holloway remedy.

The disposition in Wood therefore raises no doubt about the consistency of the Wood Court. Contrary to the majori*202ty’s conclusion, see ante, at 168-169, n. 2, there was no tension at all between acknowledging the rule of reversal to be applied when a judge fails to enquire into a known risk of prospective conflict, Wood, 450 U. S., at 272, n. 18, while at the same time sending the Wood case itself back for a determination about actual, past conflict, id., at 273-274. Wood simply followed and confirmed the pre-existing scheme established by Holloway and Sullivan. When a risk of conflict appears before a proceeding has been held or completed and a judge fails to make a prospective enquiry, the remedy is to vacate any subsequent judgment against the defendant. See Holloway, 435 U. S., at 491. When the possibility of conflict does not appear until a proceeding is over and any enquiry must be retrospective, a defendant must show actual conflict with adverse effect. See Sullivan, supra, at 349.

Wood, then, does not affect the conclusion that would be reached here on the basis of Holloway and Sullivan. This case comes to us with the finding that the judge who appointed Saunders knew or should have known of the risk that he would be conflicted owing to his prior appointment to represent the victim of the crime, 74 F. Supp. 2d, at 613-615; see n. 1, supra. We should, therefore, follow the law settled until today, in vacating the conviction and affording Mickens a new trial.

II

Since the majority will not leave the law as it is, however, the question is whether there is any merit in the rule it now adopts, of treating breaches of a judge’s duty to enquire into prospective conflicts differently depending on whether defense counsel explicitly objected. There is not. The distinction is irrational on its face, it creates a scheme of incentives to judicial vigilance that is weakest in those cases presenting the greatest risk of conflict and unfair trial, and it reduces the so-called judicial duty to enquire into so many empty words.

*203The most obvious reason to reject the majority’s rule starts with the accepted view that a trial judge placed on notice of a risk of prospective conflict has an obligation then and there to do something about it, Holloway, supra, at 484. The majority does not expressly repudiate that duty, see ante, at 167-168, which is too clear for cavil. It should go without saying that the best time to deal with a known threat to the basic guarantee of fair trial is before the trial has proceeded to become unfair. See Holloway, supra, at 484; Glasser, 315 U. S., at 76. Cf. Pate v. Robinson, 383 U. S. 375, 386-387 (1966) (judge’s duty to conduct hearing as to competency to stand trial). It would be absurd, after all, to suggest that a judge should sit quiescent in the face of an apparent risk that a lawyer’s conflict will render representation illusory and the formal trial a waste of time, emotion, and a good deal of public money. And as if that were not bad enough, a fajlure to act early raises the specter, confronted by the Holloway Court, that failures on the part of conflicted counsel will elude demonstration after the fact, simply because they so often consist of what did not happen. 435 U. S., at 490-492. While a defendant can fairly be saddled with the characteristically difficult burden of proving adverse effects of conflicted decisions after the fact when the judicial system was not to blame in tolerating the risk of conflict, the burden is indefensible when a judge was on notice of the risk but did nothing.

With so much at stake, why should it matter how a judge learns whatever it is that would point out the risk to anyone paying attention? Of course an objection from a conscientious lawyer suffices to put a court on notice, as it did in Holloway; and probably in the run of multiple-representation cases nothing short of objection will raise the specter of trouble. But sometimes a wide-awake judge will not need any formal objection to see a risk of conflict, as the federal habeas court’s finding in this very case shows. 74 F. Supp. 2d, at 613-615. Why, then, pretend contrary to fact *204that a judge can never perceive a risk unless a lawyer points it out? Why excuse a judge’s breach of judicial duty just because a lawyer has fallen down in his own ethics or is short on competence? Transforming the factually sufficient trigger of a formal objection into a legal necessity for responding to any breach of judicial duty is irrational.

Nor is that irrationality mitigated by the Government’s effort to analogize the majority’s objection requirement to the general rule that in the absence of plain error litigants get no relief from error without objection. The Government as amicus argues for making a formal objection crucial because judges are not the only ones obliged to take care for the integrity of the system; defendants and their counsel need inducements to help the courts with timely warnings. Brief for United States as Amicus Curiae 9, 26-27. The fallacy of the Government’s argument, however, has been on the books since Wood was decided. See 450 U. S., at 265, n. 5 (“It is unlikely that [the lawyer on whom the conflict of interest charge focused] would concede that he had continued improperly to act as counsel”). The objection requirement works elsewhere because the objecting lawyer believes that he sights an error being committed by the judge or opposing counsel. See, e. g., United States v. Vonn, ante, at 72-73 (error in judge’s Rule 11 plea colloquy). That is hardly the motive to depend on when the risk of error, if there is one, is being created by the lawyer himself in acting subject to a risk of conflict, 227 F. 3d 203, 213-217 (CA4 2000), vacated en banc, 240 F. 3d 348 (CA4 2001). The law on conflicted counsel has to face the fact that one of our leading cases arose after a trial in which counsel may well have kept silent about conflicts not out of obtuseness or inattention, but for the sake of deliberately favoring a third party’s interest over the clients, and this very case comes to us with reason to suspect that Saunders suppressed his conflicts for the sake of a second fee in a case getting public attention. While the perceptive and conscientious lawyer (as in Holloway) needs *205nothing more than ethical duty to induce an objection, the venal lawyer is not apt to be reformed by a general rule that says his client will have an easier time reversing a conviction down the road if the lawyer calls attention to his own venality.10

The irrationality of taxing defendants with a heavier burden for silent lawyers naturally produces an equally irrational scheme of incentives operating on the judges. The judge’s duty independent of objection, as described in Sullivan and Wood, is made concrete by reversal for failure to honor it. The plain fact is that the specter of reversal for failure to enquire into risk is an incentive to trial judges to keep their eyes peeled for lawyers who wittingly or otherwise play loose with loyalty to their clients and the fundamental guarantee of a fair trial. See Wheat, 486 U. S., at 161. Cf. Pate, supra, at 386-387 (reversal as remedy for state trial judge’s failure to discharge duty to ensure competency to stand trial). That incentive is needed least when defense counsel points out the risk with a formal objection, *206and needed most with the lawyer who keeps risk to himself, quite possibly out of self-interest. Under the majority’s rule, however, it is precisely in the latter situation that the judge’s incentive to take care is at its ebb. With no objection on record, a convicted defendant can get no relief without showing adverse effect, minimizing the possibility of a later reversal and the consequent inducement to judicial care.11 This makes no sense.

The Court’s rule makes no sense unless, that is, the real point of this case is to eliminate the judge’s constitutional duty entirely in no-objection cases, for that is certainly the practical consequence of today’s holding. The defendant has the same burden to prove adverse effect (and the prospect of reversal is the same) whether the judge has no reason to know of any risk or every reason to know about it short of *207explicit objection.12 In that latter case, the duty explicitly described in Sullivan and Wood becomes just a matter of words, devoid of sanction; it ceases to be any duty at all.

As that duty vanishes, so does the sensible regime under which a defendant’s burden on conflict claims took account of the opportunities to ensure against conflicted counsel in the first place. Convicted defendants had two alternative avenues to show entitlement to relief. A defendant might, first, point to facts indicating that a judge knew or should have known of a “ ‘particular conflict,’ ” Wood, 450 U. S., at 272, n. 18 (quoting Sullivan, 446 U. S., at 347), before that risk had a chance to play itself out with an adverse result. If he could not carry the burden to show that the trial judge had fallen down in the duty to guard against conflicts prospec*208tively, the defendant was required to show, from the perspective of an observer looking back after the allegedly conflicted representation, that there was an actual conflict of interests with an adverse effect. The first route was preventive, meant to avoid the waste of costly after-the-fact litigation where the risk was clear and easily avoidable by a reasonably vigilant trial judge; the second was retrospective, with a markedly heavier burden justified when the judiciary was not at fault, but at least alleviated by dispensing with any need to show prejudice. Today, the former system has been skewed against recognizing judicial responsibility. The judge’s duty applies only when a Holloway objection fails to induce a resolutely obdurate judge to take action upon the explicit complaint of a lawyer facing impossible demands. In place of the forsaken judicial obligation, we can expect more time-consuming post-trial litigation like this, and if this case is any guide, the added time and expense are unlikely to purchase much confidence in the judicial system.13

I respectfully dissent.

dissenting.

The Commonwealth of Virginia seeks to put the petitioner, Walter Mickens, Jr., to death after having appointed to represent him as his counsel a lawyer who, at the time of the murder, was representing the very person Mickens was accused of killing. I believe that, in a case such as this one, a categorical approach is warranted and automatic reversal is required. To put the matter in language this Court has previously used: By appointing this lawyer to represent Mick-ens, the Commonwealth created a “structural defect affecting the framework within which the trial [and sentencing] proceeds, rather than simply an error in the trial process itself.” Arizona v. Fulminante, 499 U. S. 279, 310 (1991).

The parties spend a great deal of time disputing how this Court’s precedents of Holloway v. Arkansas, 435 U. S. 475 (1978), Cuyler v. Sullivan, 446 U. S. 335 (1980), and Wood v. Georgia, 450 U. S. 261 (1981), resolve the case. Those precedents involve the significance of a trial judge’s “failure to inquire” if that judge “knew or should have known” of a “potential” conflict. The majority and dissenting opinions dispute the meaning of these cases as well. Although I express no view at this time about how our precedents should treat most ineffective-assistance-of-counsel claims involving an alleged conflict of interest (or, for that matter, whether Holloway, Sullivan, and Wood provide a sensible or coherent framework for dealing with those cases at all), I am convinced that this case is not governed by those precedents, for the following reasons.

*210First, this is the kind of representational incompatibility that is egregious on its face. Mickens was represented by the murder victim’s lawyer; that lawyer had represented the victim on a criminal matter; and that lawyer’s representation of the victim had continued until one business day before the lawyer was appointed to represent the defendant.

Second, the conflict is exacerbated by the fact that it occurred in a capital murder case. In a capital case, the evidence submitted by both sides regarding the victim’s character may easily tip the scale of the jury’s choice between life or death. Yet even with extensive investigation in post-trial proceedings, it will often prove difficult, if not impossible, to determine whether the prior representation affected defense counsel’s decisions regarding, for example: which avenues to take when investigating the victim’s background; which witnesses to call; what type of impeachment to undertake; which arguments to make to the jury; what language to use to characterize the victim; and, as a general matter, what basic strategy to adopt at the sentencing stage. Given the subtle forms that prejudice might take, the consequent difficulty of proving actual prejudice, and the significant likelihood that it will nonetheless occur when the same lawyer represents both accused killer and victim, the cost of litigating the existence of actual prejudice in a particular case cannot be easily justified. Cf. United States v. Cronic, 466 U. S. 648, 657-658 (1984) (explaining the need for categorical approach in the event of “actual breakdown of the adversarial process”).

Third, the Commonwealth itself created the conflict in the first place. Indeed, it was the same judge who dismissed the case against the victim who then appointed the victim’s lawyer to represent Mickens one business day later. In light of the judge’s active role in bringing about the incompatible representation, I am not sure why the concept of a judge’s “duty to inquire” is thought to be central to this case. No “inquiry” by the trial judge could have shed more light *211on the conflict than was obvious on the face of the matter, namely, that the lawyer who would represent Mickens today is the same lawyer who yesterday represented Mickens’ alleged victim in a criminal case.

This kind of breakdown in the criminal justice system creates, at a minimum, the appearance that the proceeding will not “‘reliably serve its function as a vehicle for determination of guilt or innocence,’ ” and the resulting “ ‘criminal punishment’” will not “‘be regarded as fundamentally fair.’” Fulminante, supra, at 310. This appearance, together with the likelihood of prejudice in the typical case, is serious enough to warrant a categorical rule — a rule that does not require proof of prejudice in the individual case.

The Commonwealth complains that this argument “relies heavily on the immediate visceral impact of learning that a lawyer previously represented the victim of his current client.” Brief for Respondent 34. And that is so. The “visceral impact,” however, arises out of the obvious, unusual nature of the conflict. It arises from the fact that the Commonwealth seeks to execute a defendant, having provided that defendant with a lawyer who, only yesterday, represented the victim. In my view, to carry out a death sentence so obtained would invariably “diminis[h] faith” in the fairness and integrity of our criminal justice system. Young v. United States ex rel. Vuitton et Fils S. A., 481 U. S. 787, 811-812 (1987) (plurality opinion). Cf. United States v. Olano, 507 U. S. 725, 736 (1993) (need to correct errors that seriously affect the “ ‘fairness, integrity or public reputation of judicial proceedings’ ”). That is to say, it would diminish that public confidence in the criminal justice system upon which the successful functioning of that system continues to depend.

I therefore dissent.