The Co-Defendant Confession Problem (Bruton) in a Texas Joint Trial

The Bruton Problem: A Confession You Cannot Cross-Examine

The Sixth Amendment secures to an accused the right to confront and cross-examine the witnesses against him. That right collides with joint trials when the State offers a non-testifying co-defendant’s confession that also incriminates the defendant sitting beside him. The confessing co-defendant does not take the stand, so the defendant can never cross-examine the person whose words help convict him.

In Bruton v. United States, the Supreme Court held that admitting such a confession at a joint trial violates the implicated defendant’s confrontation right, even when the judge instructs the jury to consider the confession only against the co-defendant who made it. Two men were tried together for robbery; a postal inspector recounted one defendant’s oral confession that named the other as his accomplice. The Court reversed, reasoning that a co-defendant’s accusatory confession is both powerfully incriminating and inherently suspect, because an accomplice has every reason to shift blame, and its unreliability is compounded when the accomplice cannot be tested by cross-examination.

A key point frames everything that follows: the co-defendant’s confession is ordinarily hearsay as to the other defendant and is not independently admissible against him. It comes in only against the declarant, with a limiting instruction. Bruton holds that in this narrow context the instruction is not enough, and the Constitution demands a different solution. The sections below explain why the instruction fails, how the rule survives even when the defendant has confessed too, and what a Texas court must do to fix the problem.

It is also worth locating the rule’s limits at the outset. Bruton addresses a specific evil — a co-defendant’s accusing confession that the defendant cannot cross-examine — not a general objection to damaging evidence. It applies when the statement is offered through a third party, such as an officer recounting what the co-defendant said, and when the co-defendant will not take the stand. Understanding that scope keeps the objection sharp and helps counsel separate the confessions that genuinely trigger the rule from those that do not.

Why a Limiting Instruction Is Not Enough

The ordinary assumption of the trial system is that jurors follow instructions, including an instruction to weigh a piece of evidence against one defendant but not another. Bruton carves out a narrow exception to that assumption because some evidence is too damaging for the human mind to compartmentalize.

The Court explained that there are 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 limitations of the jury system cannot be ignored. A co-defendant’s confession that expressly names the defendant as a participant is exactly that kind of evidence. Telling jurors to use a confession against the man who gave it, while ignoring its plain accusation of the man next to him, asks for a mental act the Court concluded jurors cannot reliably perform. The confession is, in the Court’s words, deliberately spread before the jury, and its accusing content cannot be unheard.

This is why a Bruton problem cannot be waved away with a jury charge. When a facially incriminating co-defendant confession is in play, the defense should not accept a limiting instruction as a cure. The correct response is to force a structural fix — separate the trials, keep the confession out, or transform it so it no longer accuses the defendant — before the jury ever hears the accusing words. Everything turns on which of those fixes the court adopts, and the confession’s exact wording drives that choice.

The point about wording deserves emphasis. Whether a confession creates a Bruton problem, and which remedy fits, both turn on the precise words the jury will hear. A statement that names the defendant is the core violation; a statement carefully rewritten may be admissible; a statement with an obvious hole where the name was is not. Counsel should therefore insist on seeing the exact form of any co-defendant statement the State intends to use, well before it is read to the jury, so the objection can be litigated on the actual text rather than a summary.

Interlocking Confessions Do Not Save the Joint Trial

Prosecutors sometimes argue that Bruton should not apply when the defendant has confessed too, and his confession “interlocks” with the co-defendant’s — that is, both statements describe the same events. The Supreme Court rejected that argument, so a defendant’s own confession does not open the door to his co-defendant’s.

In Cruz v. New York, the Court held that where a non-testifying co-defendant’s confession incriminating the defendant is not directly admissible against him, the Confrontation Clause bars its admission at the joint trial even if the jury is instructed to disregard it, and even if the defendant’s own interlocking confession is admitted against him. The Court explained that the interlocking quality of the two statements goes to reliability, not to harmfulness. If anything, a matching co-defendant confession can be more devastating, because it corroborates the very statement the defendant may be trying to challenge as coerced or false. The jury’s inability to follow a limiting instruction does not improve just because the defendant also spoke.

The practical takeaway is that counsel should not concede a Bruton claim merely because the client gave a statement. The existence of the client’s own confession may bear on whether any error was harmless on appeal, or on whether the co-defendant’s statement is independently reliable enough to admit under a hearsay exception, but it does not neutralize the confrontation violation at trial. The co-defendant’s accusing confession still cannot be placed before a joint jury with only an instruction to look away.

There is a strategic dimension too. A defendant who is contesting the voluntariness or accuracy of his own statement has the most to lose from an interlocking co-defendant confession, because the second statement lends the first a false aura of corroboration in the jurors’ minds. Cruz recognizes that dynamic, and it is why counsel challenging a client’s confession should be especially alert to a co-defendant’s parallel statement and should press for severance or exclusion rather than accept the two statements side by side before the same jury.

Redaction: What Cures the Problem and What Does Not

Severance is not the only fix. A confession can sometimes be edited so that it no longer accuses the defendant, which lets the joint trial proceed. But the Supreme Court has drawn fine lines around what redaction actually works, and a clumsy edit can leave the violation fully intact.

In Richardson v. Marsh, the Court held that the Confrontation Clause is not violated when a co-defendant’s confession is redacted to eliminate not only the defendant’s name but any reference to his existence, and a proper limiting instruction is given — even if other evidence later links the defendant to the statement. That is inferential incrimination, which the Court treats as different from a confession that accuses on its face. In Gray v. Maryland, by contrast, the Court held that an obvious redaction — a blank space, the word “deleted,” or a similar symbol where the name used to be — still violates Bruton, because such a gap points directly at the defendant and invites the jury to fill in the obvious answer. The table summarizes where each method lands.

Because the difference between a lawful and an unlawful redaction can come down to a single word or an obvious gap, defense counsel should scrutinize any proposed edited confession line by line and object when the redaction still lets the jury see that a name was removed.

The line between Richardson and Gray is functional, not formal. What matters is whether the edited statement, as the jury actually receives it, points at the defendant. A confession that reads naturally as a narrative about unnamed others can be permissible; one that visibly announces an omission invites the jury to supply the missing name and defeats the purpose of redaction. Counsel should review not only the written exhibit but how it will be read aloud, since inflection and a follow-up question can reintroduce the very reference the redaction was supposed to remove.

Samia and Neutral Placeholders

The Supreme Court refined the redaction line again in 2023, and the refinement cuts against defendants. Not every substitute for a name triggers Bruton; a neutral placeholder that does not announce a deletion can be admissible with a limiting instruction.

In Samia v. United States, a co-defendant’s confession was altered so that it referred to the defendant only as “the other person” rather than by name or with a blank. The Court held that the Confrontation Clause was not violated by admitting a confession that did not directly inculpate the defendant and was accompanied by a proper limiting instruction. The Court distinguished Gray: a neutral reference to “another person” is not the same as an obvious blank or the word “deleted,” and it does not, on its face, point to the defendant. The Court also stressed that the Confrontation Clause does not provide a freestanding protection against the risk that a jury might, through inference from other evidence, connect a redacted confession to the defendant.

For Texas practice, Samia means the fight over a co-defendant’s confession has become a fight over exact wording and context. A rewrite that neutrally describes an unnamed other participant may survive, while an edit that leaves an obvious hole will not. Counsel should probe whether the placeholder is truly neutral, whether the surrounding testimony and the prosecutor’s argument effectively re-insert the client’s identity, and whether, given the number of participants, “the other person” can only mean the defendant. Where the placeholder still functions as an accusation, the Bruton objection and the severance request remain live.

Samia should not be read as an invitation to launder any accusation through a placeholder. The Court’s holding rested on a confession that read as a neutral account and on a limiting instruction, in a multi-participant case where “the other person” did not unmistakably mean the defendant. Where the surrounding evidence, the number of participants, or the prosecutor’s argument makes the placeholder transparent, the statement moves back toward Gray. The defense role is to hold the State to a truly neutral rewrite and to build a record when it falls short.

The Remedies in a Texas Joint Trial

When a co-defendant’s confession creates a confrontation problem, a Texas trial court has three ways to resolve it, and the defense should be ready to argue for the one that most protects the client. The confession is generally inadmissible hearsay as to the non-declarant defendant, so the only question is how to keep its accusing content away from a jury that is judging that defendant.

First, the court can sever the trials under article 36.09. A non-testifying co-defendant’s confession that implicates the movant is a classic showing that a joint trial would be prejudicial, and separate trials eliminate the problem because the confession simply is not offered in the implicated defendant’s case. Second, the court can exclude the confession from the joint trial entirely, or require the State to elect not to use it. Third, the court can order a lawful redaction or rewrite that, under Richardson and Samia, removes any facial reference to the defendant — but not the kind of obvious deletion condemned in Gray. Whether to sever, exclude, or redact is committed to the trial court, which is why the defense should propose the specific remedy it wants and object on the record to any partial fix.

One situation dissolves the problem altogether: if the confessing co-defendant actually takes the witness stand and is subject to cross-examination, there is no Bruton violation, because the defendant can then confront the person who accused him. A related note is that Bruton concerns testimonial statements offered for their truth; a co-defendant’s confession recounted by an investigator is the paradigm. For the broader framework of proving that a joint trial would be prejudicial, and for conflicting-defense theories, see the companion pages under this motion.

Sequencing these remedies is itself a tactical decision. A defendant who prefers a joint trial for strategic reasons may push for exclusion or a neutral rewrite rather than severance, while a defendant who wants to be tried alone will argue that no redaction can cure the problem and that separation is the only adequate remedy. Because the choice belongs to the trial court, the defense should articulate which remedy it wants, explain why lesser measures are inadequate for this particular confession, and preserve the objection if the court adopts a fix that leaves the accusation intact.

Where This Fits

This guide is one of four situations that arise under the Motion to Sever Defendants. 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 co-defendant's confession that implicates the defendant at a joint trial violates the Confrontation Clause despite a limiting instruction.
• Cruz v. New York, 481 U.S. 186 (1987) — The Confrontation Clause bars a non-testifying co-defendant's incriminating confession even when it interlocks with the defendant's own confession.
• Richardson v. Marsh, 481 U.S. 200 (1987) — A confession redacted to eliminate the defendant's name and any reference to his existence, with a limiting instruction, does not violate confrontation.
• Gray v. Maryland, 523 U.S. 185 (1998) — Replacing the defendant's name with an obvious blank, the word deleted, or a symbol still violates Bruton because it points directly at the defendant.
• Samia v. United States, 599 U.S. 635 (2023) — A confession redacted to a neutral placeholder that does not directly inculpate the defendant, with a limiting instruction, does not violate confrontation.

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