Dissenting.
I fully join the Majority in finding improper the prosecutor’s closing argument revealing defendant, Terry Brown (“Appellee”), as the “guy” in the previously redacted statement of Miguel Garcia, his non-testifying co-defendant. Maj. Op. at 398-99, 925 A.2d at 159-60. Respectfully, I diverge from the Majority premised upon my conviction that this type of error is never amenable to curative instructions. Rather than viewing this as a category distinct from the typical violation of Bruton v. United States, 391 U.S. 123, 88 S.Ct. 1620, 20 *405L.Ed.2d 476 (1968), I conclude that a prosecutor’s unmasking of a defendant’s redacted identity is a sub-category of Bruton violation, subject to Bruton’s per se rule that jury instructions cannot remedy the “spillover” prejudice resulting from a non-testifying co-defendant’s statement. Because I conclude that Bruton applies to the prosecutorial references here in question, it would remain for me only to determine whether the error in question was harmless. As discussed below, I conclude that the violation was not harmless and, consequently, I would affirm the Superior Court’s determination that the Bruton violation necessitates a remand for a new trial.
A defendant “is deprived of his rights under the Confrontation Clause when his non-testifying codefendant’s confession naming him as a participant in the crime is introduced at their joint trial, even if the jury is instructed to consider that confession only against the codefendant.” Richardson v. Marsh, 481 U.S. 200, 201-02, 107 S.Ct. 1702, 95 L.Ed.2d 176 (1987). In Bruton, the United States Supreme Court explained why no jury instruction can cure this type of prejudice:
[T]here are some contexts in which the risk that the jury will not, or cannot, follow instructions is so great, and the consequences of failure so vital to the defendant, that the practical and human limitations of the jury system cannot be ignored. Such a context is presented here, where the powerfully incriminating extrajudicial statements of a codefendant, who stands accused side-by-side with the defendant, are deliberately spread before the jury in a joint trial. Not only are the incriminations devastating to the defendant but their credibility is inevitably suspect, a fact recognized when accomplices do take the stand and the jury is instructed to weigh their testimony carefully given the recognized motivation to shift blame onto others. The unreliability of such evidence is intolerably compounded when the alleged accomplice, as here, does not testify and cannot be tested by cross-examination. It was against such threats to a fair trial that the Confrontation Clause was directed.
*406Bruton, 891 U.S. at 135-136, 88 S.Ct. 1620 (internal citations omitted).
For the same reasons, I believe that a defendant is equally deprived of his rights under the Confrontation Clause if an otherwise effective redaction is corrupted by a prosecutor’s comment. The jury is no less aware of the import of the co-defendant’s implication of the defendant, and the defendant is no more able to cross-examine the co-defendant than in a more conventional Bruton scenario. As with the typical Bruton violation, we cannot assume that a jury will be able to perform the necessary “mental gymnastics” of limiting its consideration of an incriminating statement as against the confessing co-defendant and not against the defendant following the improper revelation of his previously redacted identity. Frazier v. Cupp, 394 U.S. 731, 735, 89 S.Ct. 1420, 22 L.Ed.2d 684 (1969). Accordingly, I find no justification for a divergent analysis for the undoing of a redaction than that applied to an originally improper redaction.
The Majority reaches a different conclusion, based in part on the decision in Commonwealth v. McCrae, 574 Pa. 594, 832 A.2d 1026 (2003), which I find distinguishable. In that case involving a joint trial of two defendants, a witness testified at a preliminary hearing and implicated both of the defendants. However, at the time of the preliminary hearing, only one of the defendants had been charged, and thus, only that defendant had the opportunity to cross-examine the witness. By the time of the defendants’ joint trial, however, the witness had become unavailable. To address the situation, the trial court admitted the witness’s statement only after it was redacted to remove references to the later-charged defendant, who had no opportunity to cross-examine the witness. Notwithstanding the redaction, the prosecutor in his opening statement and closing argument linked the witness’s statement to the defendant whose name had been redacted. This Court found no Bruton violation had occurred, however, because the statement in that case was that of a witness to a crime and not a co-defendant.
*407While the Majority emphasizes the fact that the alleged violation in McCrae, as in the case at bar, came in the form of a prosecutor’s statements undoing the trial court’s redaction, the analysis in McCrae focused on the identity of the statement’s author, rather than how the redacted information was revealed:
Appellant’s claim ... pivots on the faulty presumption that the Bruton rule applies to [the witness’s] testimony. It does not. The statement here was not made by [a co-defendant]. Moreover, [the witness] was never charged in relation to the criminal incident, let alone tried jointly with appellant. It is the particularly “devastating” prejudicial effect and inherent “unreliability” of a directly incriminating statement made by a non-testifying co-defendant that powered Bruton’s exception to the general rule that cautionary charges are enough to avoid spillover prejudice in joint trials. [Bruton, 391 U.S. at 136, 88 S.Ct. 1620.] These concerns are simply not present in the instant context involving the statement of a non-party eyewitness to the crime.
Id. at 1038 (emphasis in original). The Court therefore concluded, “Bruton is inapplicable to statements made by an individual other than a non-testifying co-defendant at a joint trial of co-defendants.” Id.1 Because the case at bar involves a Bru,ton-governed statement of a co-defendant rather than a *408witness, I conclude that our decision in McCrae is distinguishable.
Similarly, unlike the Majority, I find the decision in Frazier v. Cupp distinguishable. In Frazier, the Court considered whether the rule in Bruton applied to a prosecutor’s opening statements. In that case, an alleged accomplice of the defendant pleaded guilty and was awaiting sentencing at the time of the defendant’s trial. The prosecutor in opening statements in the defendant’s case referred to the expected testimony of the accomplice. At trial, however, the accomplice invoked his privilege against self-incrimination on the witness stand and declined to testify. The defendant claimed that this violated Bruton because the prosecution had the benefit of placing a testimonial-like statement before the jury, without subjecting the speaker to cross-examination. The Court focused on the fact that the case did not involve a statement of a co-defendant in a joint trial, but instead related to a prosecutor’s commentary in an opening statement on what evidence might be introduced later during trial.
It may be that some remarks included in an opening or closing statement could be so prejudicial that a finding of error, or even constitutional error, would be unavoidable. But here we have no more than an objective summary of evidence which the prosecutor reasonably expected to produce. * * * Even if it is unreasonable to assume that a jury can disregard a coconspirator’s statement when introduced against one of two joint defendants, it does not seem at all remarkable to assume that the jury will ordinarily be able to limit its consideration to the evidence introduced during the trial.
Frazier, 394 U.S. at 736, 89 S.Ct. 1420 (quotation marks omitted).
Unlike in Frazier, the statement at issue in this case was a statement of a non-testifying co-defendant introduced into evidence at a joint trial of co-defendants.2 It is beyond cavil *409that the co-defendant’s statement itself falls under the protections provided by Bruton. In my mind, the form of the violation, in this case the reference in the prosecutor’s closing statement, cannot be the basis for a legal distinction when the effect is to place directly before the jury “the powerfully incriminating extrajudicial statements of a codefendant, who stands accused side-by-side with the defendant,” where such statements cannot be tested by cross-examination. Bruton, 891 U.S. at 135-36, 88 S.Ct. 1620. The violation I find in this case arises because the prosecutor’s improper reference in closing arguments effectively introduced the non-testifying co-defendant’s statement to the jury as though it had never been redacted. Respectfully, this Court should not admit a violation of the fundamental right of confrontation and cross-examination by means of a back-door revelation, when Bruton so carefully guards the front door. As the Majority aptly notes: “There is no point in redacting and sanitizing otherwise inculpatory statements of a non-testifying co-defendant, to facilitate a joint trial, if that protective measure approved by the High Court to comport with the Confrontation Clause could be deliberately and directly undone by lawyer commentary.” Maj. Op. at 398, 925 A.2d at 159. Accordingly, I maintain that a violation of the rule in Bruton occurred when the prosecutor, during her closing statement, effectively undid the Bruton-compliant redaction of the non-testifying co-defendant’s statement admitted at trial.I *3
*410When a Bruton violation has been found, the Court must consider whether the violation necessitates a new trial or merely constitutes harmless error under the circumstances in which it occurred.
An error will be deemed harmless if: (1) the error did not prejudice the defendant or the prejudice was de [minimis ]; [or] (2) the erroneously admitted evidence was merely cumulative of other untainted evidence which was substantially similar to the erroneously admitted evidence; or (3) the properly admitted and uncontradicted evidence of guilt was so overwhelming and the prejudicial effect of the error was so insignificant by comparison that the error could not have contributed to the verdict. The Commonwealth bears the burden to prove harmlessness beyond a reasonable doubt.
Commonwealth v. Markman, 916 A.2d 586, 603 (Pa.2007) (internal citations and quotation marks omitted).
In the case at bar, I cannot conclude beyond a reasonable doubt that the Bruton violation did not contribute to the verdict. While the Commonwealth relies on the testimony of the eyewitness, Anthony Cheatham, his testimony is suspect because he is a potential accomplice in the crime. Moreover, his testimony, which implicated Appellee by name, is contradicted by co-defendant Garcia’s statement, which, as redacted, did not implicate Appellee by name. The versions of events offered by Garcia and Cheatham differed as to whether Garcia or Cheatham accompanied the shooter (or Appellee in Cheat-ham’s version) during the murder.4 Additionally, at the time *411of the arrest, the murder weapon was controlled by Garcia, not Appellee. Finally, Garcia’s statement is not cumulative of other evidence because it is the only evidence presented that spans the entire period from the murder to the arrest. While there is no indication of a nefarious intent by the prosecutor to vitiate the Bruton-compliant redaction, her comments nevertheless tied Appellee by name to the “other guy” referenced in Garcia’s statement. While other incriminating evidence existed, Garcia’s statement was important evidence placing Appellee alone with Garcia after the murder with the gun in Appellee’s hand and corroborating Cheatham’s testimony that Appellee was the shooter. Accordingly, I find a reasonable possibility that the error may have contributed to the verdict such that I cannot conclude that the error was harmless. Thus, I would affirm the Superior Court’s decision to remand for a new trial.
Chief Justice CAPPY joins this Dissenting Opinion.. Notably, the Court was not united in its conclusion that the undoing of a witness’s redacted statement was categorically different from typical Bruton violations.
On the broader legal points involved, while I agree with the Majority that the United States Supreme Court's specific holding in [Bruton] is not controlling here, I would not apply as categorical an approach to the Sixth Amendment principles underlying Bruton as the Majority seems to apply. In particular, I would not exclude the possibility that such principles might render cautionaiy instructions insufficient to prevent the possibility of spillover prejudice in some larger set of circumstances analogous to, but not overlapping with, those before the United States Supreme Court in Bruton.
McCrae, 832 A.2d at 1041 (Saylor, J., concurring, joined by Cappy, C.J., and Nigro, J.). The concurring justices, however, agreed that any error committed was harmless.
. Additionally, I find the Majority's reliance on United States v. Sandini, 888 F.2d 300 (3d Cir.1989), and United States v. Espinosa, 771 F.2d 1382 (10th Cir.1985), untenable because neither involves an improper *409identification of the defendant by name in connection with a previously, properly redacted and admitted statement of a non-testifying co-defendant.
. Although the discussion in the Majority Opinion and the arguments of the parties focus upon the prosecutor’s reference to Appellee by name in connection with Garcia's statement at page 103 of the Notes of Testimony, minutes later the prosecutor again tied Appellee, by name, to Garcia's statement:
I submit to you that there wouldn’t have been any fingerprints once that gun was in the snow and with all that tape around it, it got wet in the snow, the officers handled it to take it apart, the defendant’s [sic] both handled it, so what fingerprints are you going to get on the tape? It’s an old ugly gun, but it worked. It killed Maty Edmond. It did just what Antonio Lambert [Appellee] wanted it to do. He managed to shoot her. Why? Because she didn’t give up her pocket*410book. She resisted it. And that’s exactly what Mr. Garcia said in his statement.
N.T., 6/7/02, at 106 (emphasis added). While there is no clear objection by defense counsel to this comment, the trial court had permitted argument to "go ahead” after the first comment without discussion of the objection. Id. at 104. Moreover, I find this second incident relevant to the harmless error analysis as it compounds the original Bruton violation.
. The Majority recites the facts based on Cheatham's testimony without acknowledging the discrepancy between Cheatham and Garcia’s versions of the events as described fully in the Superior Court’s published decision, Commonwealth v. Brown, 853 A.2d 1029, 1031-1033 (Pa.Super.2004).