Bruton and the Non-Testifying Codefendant's Confession

What Is the Bruton Problem in a Federal Joint Trial?

When two people are tried together in federal court, the Bruton rule bars the government from introducing a non-testifying codefendant’s confession that names and incriminates the other defendant. Because the confessing codefendant does not take the stand, the accused cannot cross-examine him, and the Sixth Amendment forbids it.

The rule comes from Bruton v. United States, 391 U.S. 123 (1968). A postal inspector testified that Bruton’s codefendant, Evans, had orally confessed that he and Bruton committed an armed robbery. Evans did not testify, so Bruton could not question him about the accusation. The trial judge told the jury to consider the confession only against Evans, not Bruton. The Supreme Court held that this instruction was not enough: admitting the confession violated Bruton’s “right of cross-examination secured by the Confrontation Clause of the Sixth Amendment,” and it overruled the earlier contrary rule of Delli Paoli.

The Confrontation Clause guarantees a defendant the chance to test accusations by cross-examining the witnesses against him. A codefendant’s out-of-court confession is a classic accusation the defense cannot test when the codefendant stays silent at trial. That is the heart of the problem: a “powerfully incriminating” statement reaches the jury through the back door, spoken by someone the defendant can never question. The joint trial, meant to promote efficiency, becomes the mechanism that strips away a core constitutional protection.

It is important to see how narrow this exception is. The ordinary rule in a joint trial is the opposite: a witness whose testimony is admitted against one defendant is not treated as a witness “against” a codefendant when the jury is told to consider it only against the first defendant, and courts presume jurors follow that instruction. Bruton carves out a single, tightly drawn situation — a facially incriminating confession by a codefendant who will not testify — where that presumption cannot be trusted. Everything in the case law that follows is an effort to police the boundary of that narrow exception, which is why the exact wording of the confession matters so much.

Why a Limiting Instruction Cannot Cure It

Prosecutors often argue that a limiting instruction telling jurors to weigh the confession only against the person who made it removes the harm. Bruton rejected that. Some evidence is too devastating for a jury to disregard, and the confession of a codefendant pointing at the accused sits at the top of that list.

The Bruton Court explained that “in the context of a joint trial we cannot accept limiting instructions as an adequate substitute for petitioner’s constitutional right of cross-examination,” because the effect “is the same as if there had been no instruction at all.” A jury cannot realistically hear one defendant admit that both of them robbed a bank and then wall that admission off when judging the second defendant. The risk that jurors will look to the confession anyway is simply too great.

The Court reinforced this in Cruz v. New York, 481 U.S. 186 (1987), which closed a tempting loophole. The government argued that when the defendant has given his own confession that “interlocks” with the codefendant’s, the codefendant’s statement does no additional damage. The Court disagreed, holding that the Confrontation Clause bars the non-testifying codefendant’s incriminating confession at a joint trial “even if the jury is instructed not to consider it against the defendant, and even if the defendant’s own confession is admitted against him.” The existence of a second confession does not rescue the first.

The reason the Court treats these confessions differently comes down to what it called the “practical and human limitations of the jury system.” Most improperly admitted evidence can be cured by an instruction, and Bruton did not disturb that general principle. But a codefendant’s confession that a defendant helped commit the crime is, in the Court’s words, “powerfully incriminating,” and it is spread before the jury by someone the defense cannot question. When the risk that jurors will disregard the instruction is high and the stakes for the defendant are that serious, the Court concluded, the ordinary faith in limiting instructions has to yield to the confrontation right.

Redaction: What Works and What Fails

Rather than lose the joint trial, the government usually offers to redact the confession. Redaction can cure the problem, but only if it is done thoroughly. Editing that leaves the accusation obvious does nothing; editing that removes the defendant entirely from the statement can be enough.

In Richardson v. Marsh, 481 U.S. 200 (1987), the Court held that “the Confrontation Clause is not violated by the admission of a nontestifying codefendant’s confession with a proper limiting instruction when . . . the confession is redacted to eliminate not only the defendant’s name, but any reference to his or her existence.” The key is that the edited statement is not incriminating on its face; it becomes damaging only when the jury links it to other evidence. A confession that must be connected to the defendant by inference falls outside Bruton’s narrow rule.

Gray v. Maryland, 523 U.S. 185 (1998), drew the other boundary. The prosecution swapped the defendant’s name for a blank space and the word “deleted.” The Court held that redactions replacing a name with “an obvious blank space, the word ‘deleted,’ a symbol, or similarly obvious indications of alteration” still fall within Bruton’s protection, because a jury will realize the confession refers to the defendant sitting at counsel table. Such edits are “directly accusatory” and are treated like the original. The table below sorts the common approaches.

The Samia Rule: Neutral Placeholders After 2023

The most recent word from the Supreme Court came in Samia v. United States, 599 U.S. 635 (2023). It tells courts that a confession rewritten to refer to a generic “other person,” paired with a limiting instruction, generally passes constitutional muster because it does not point at the defendant on its face.

In Samia, a codefendant’s confession was altered so that references to Samia became “another person” or “someone he met.” The Court held that “the Confrontation Clause was not violated by the admission of a nontestifying codefendant’s confession that did not directly inculpate the defendant and was subject to a proper limiting instruction.” It restated the governing line: Bruton reaches only “directly accusatory” statements, not those that “become incriminating only when linked with evidence introduced later at trial.”

Samia matters because it approves a technique — rewriting rather than merely blacking out — that keeps many joint trials intact. But it is not a blank check. The Court left standing the rule that a placeholder so distinctive it obviously means the defendant — the classic example being a description “as unique as the ‘red-haired, bearded, one-eyed man-with-a-limp’” — still runs afoul of Gray. And a prosecutor who repairs the facial problem cannot then undo it by arguing in closing that the “other person” is the defendant. Where the altered confession still telegraphs the accused, the objection survives.

In day-to-day litigation, Samia turns the fight into a close reading of the edited statement. The defense examines whether the substituted phrasing — “another person,” “someone,” or “a person he knew” — genuinely conceals the defendant or merely gestures at him. If only one other person sits at the defense table, or if the surrounding facts make the placeholder’s meaning obvious, the redaction may collapse back into Gray territory despite Samia’s general approval of the technique. Counsel should also insist that the government not undo a proper redaction by inviting the jury, through argument or a pointed question, to fill in the blank with the defendant’s name.

Three Remedies: Sever, Redact, or Exclude

The fix for a Bruton problem is not automatic severance. Rule 14(a) gives the district court a menu: it may “order separate trials of counts, sever the defendants’ trials, or provide any other relief that justice requires.” A Bruton violation can be avoided in more than one way.

Three options recur. First, the court can sever, giving the accused a separate trial where the codefendant’s confession is simply irrelevant. Second, the court can require the government to redact the statement so that it no longer names or points to the defendant, following Richardson and Samia. Third, the court can bar the confession entirely, forcing the government to prove its case without it. Because the government would rather keep both its confession and its joint trial, it frequently proposes a redaction, and the fight becomes whether the proposed edit is thorough enough to satisfy Gray.

Rule 14 also equips the judge to make that call before trial. It provides that “before ruling on a defendant’s motion to sever, the court may order an attorney for the government to deliver to the court for in camera inspection any defendant’s statement that the government intends to use as evidence.” That lets the court read the actual confession, test the proposed redaction against the Bruton line, and choose the remedy before the jury ever hears a word. The prosecution proposes first, but the court decides.

Courts do not order severance lightly, and understanding why explains the preference for redaction. The Supreme Court has emphasized that joint trials serve important interests — conserving resources, sparing witnesses from testifying twice, and avoiding the risk of inconsistent verdicts. A Bruton problem does not erase those interests; it simply requires the court to protect the confrontation right by some means. That is why, when a thorough redaction under Richardson or Samia will do the job, judges frequently choose it over separate trials. The defense objective is to show that no adequate redaction is possible, so that severance or exclusion becomes the only faithful remedy.

How and When to Raise a Bruton Objection

Timing is not optional. A request to sever because of a codefendant’s confession is a motion that must be made before trial. Under Rule 12(b)(3)(D), a motion for “severance of charges or defendants under Rule 14” must be raised by pretrial motion when the basis is reasonably available, and a motion filed after the deadline is untimely absent good cause.

A well-built motion does three things. It identifies the specific statement and quotes the language that points to the client. It explains why the confession facially incriminates the defendant, or why the government’s proposed redaction leaves the accusation obvious under Gray. And it asks the court to inspect the statement in camera under Rule 14 and to choose among severance, a stricter redaction, or exclusion before opening statements. Solving the problem early is the whole point, because once the jury hears an unredacted or poorly redacted confession, an instruction to disregard it cannot reliably undo the harm.

Two practical points round out the analysis. The Bruton objection belongs to the defendant the confession points to, so counsel should be clear about whose Confrontation rights are at stake. And Bruton is one piece of a larger Confrontation Clause framework: a codefendant’s confession is generally testimonial hearsay, so the same statement may raise broader confrontation and hearsay objections apart from the Bruton rule. The defense should preserve every ground, because a joint trial that proceeds over a valid Bruton objection is a frequent and serious issue on appeal.

Preserving the objection also shapes what happens if the trial court gets it wrong. A Bruton error is not automatically fatal on appeal; reviewing courts ask whether the constitutional violation was harmless. As the Supreme Court noted in Cruz, a defendant’s own confession may be considered on appeal “in assessing whether any Confrontation Clause violation was harmless.” That makes the trial record decisive. Counsel who clearly identifies the offending statement, secures a ruling, and documents how the confession filled a gap in the government’s proof gives the appellate court the material it needs to treat the error as consequential rather than harmless.

Where This Fits

This guide is one of four situations that arise under the Motion to Sever Defendants or Counts. Start with the parent motion for the overall framework, or move to a related fact pattern:

What the Case Law Says

These decisions—verified against primary sources—control how this issue is litigated. Every case still turns on its own facts.

• Bruton v. United States, 391 U.S. 123 (1968) — Admitting a non-testifying codefendant's confession that incriminates the defendant at a joint trial violates the Confrontation Clause despite a limiting instruction.
• Richardson v. Marsh, 481 U.S. 200 (1987) — No violation where the confession is redacted to eliminate the defendant's name and any reference to the defendant's existence, with a limiting instruction.
• Gray v. Maryland, 523 U.S. 185 (1998) — Redaction replacing the defendant's name with a blank space, the word deleted, or an obvious symbol still violates Bruton.
• Samia v. United States, 599 U.S. 635 (2023) — No violation where a confession is altered to refer to another person and does not directly inculpate the defendant, with a limiting instruction.
• Cruz v. New York, 481 U.S. 186 (1987) — The Confrontation Clause bars a non-testifying codefendant's confession even when the defendant's own interlocking confession is also admitted.

General summaries of published opinions for information only — not predictions about any specific case.