Grant-of-Immunity and Other Prior Bars to Prosecution in Texas

Immunity as a Bar: Statutory, Not Constitutional

A grant of immunity is one of the few things that can defeat a Texas prosecution before any trial, but in Texas it does not come from the constitution or an inherent prosecutorial power. It is statutory, and understanding its source explains every requirement that follows.

In Graham v. State, the Court of Criminal Appeals held that in Texas the power to grant immunity from prosecution is statutory rather than constitutional, and that the State has never had a general immunity statute. Instead, the authority is derived from the statutes that let officers of the Judicial Department dismiss prosecutions — principally Code of Criminal Procedure article 32.02, which allows the attorney for the State to dismiss a case, but only with the court's consent. A grant of transactional immunity is, conceptually, a prosecutorial promise to dismiss.

That statutory root has a direct consequence: because the underlying power to dismiss requires court approval, so does a binding grant of transactional immunity. Graham made the point bluntly — when a court has not approved a prosecutor's agreement to grant immunity, there is no grant of immunity on which a defendant can rely. A defendant told he has immunity should therefore ask a hard question early: was the promise approved by a court, or is it only a prosecutor's assurance that has not yet been given legal force?

That framing also explains why immunity looks different from an ordinary favorable exercise of discretion. A prosecutor is free to decline to charge, to dismiss a weak case, or to recommend leniency, but none of those choices, standing alone, creates a legal bar the defendant can enforce against a later prosecution. Immunity is the special case in which the prosecutor's promise, once given legal force through the court's approval of a dismissal, becomes something the defendant can hold the State to. Keeping that distinction in view prevents a defendant from treating a hopeful conversation as a binding grant.

Transactional Versus Use Immunity

Not all immunity is the same, and the two principal kinds bar very different things. Transactional immunity protects the person from prosecution; use immunity protects the person's words. Confusing them is a common and costly mistake.

Transactional immunity is a promise not to prosecute for the offense or transaction covered, so it operates as a bar to the case itself. Use and derivative-use immunity is narrower. In Kastigar v. United States, the Supreme Court held that use and derivative-use immunity is coextensive with the Fifth Amendment privilege and is enough to compel testimony over a claim of the privilege — but it does not bar prosecution. Instead, it bars using the compelled testimony and anything derived from it, and it places an affirmative burden on the government, if it later prosecutes, to prove that its evidence came from a legitimate source wholly independent of the compelled testimony. The comparison is set out below.

Graham captured the Texas split neatly: a prosecutor cannot bind the State to transactional immunity without court approval, but a prosecutor may agree to use immunity — to refrain from using particular evidence — without court approval, because deciding what evidence to offer is within prosecutorial discretion.

Who Can Grant Immunity, and the Court-Approval Requirement

Because transactional immunity flows from the power to dismiss, only the actor who holds that power — the prosecutor, with the court's consent — can grant it. Promises from anyone else, or promises never blessed by a court, do not bar the case.

The authority belongs to the attorney representing the State, an officer of the Judicial Department, acting under article 32.02 with the presiding judge's approval. A promise by police, or an agreement that on the record ran only to a sheriff's department rather than the prosecuting attorney, does not create transactional immunity from prosecution; Graham turned on exactly that gap. Graham also rejected any doctrine of “equitable immunity” that would enforce an unapproved prosecutorial promise on fairness grounds alone.

Court approval is real but not onerous in form. In Smith v. State, the Court of Criminal Appeals held that as long as the judge approves the dismissal that results from an immunity agreement and is aware the dismissal is pursuant to such an agreement, the judge need not be told the agreement's specific terms for it to be enforceable. The dismissal or nolle prosequi, the court explained, is simply the final event in a transactional immunity agreement — the acknowledgment that both sides have performed. The requirement is judicial consent to the dismissal, not judicial supervision of every clause; but without that consent, there is no enforceable bar.

Enforcing the Agreement: Contract Plus Due Process

Once a valid immunity agreement exists, it is enforced with unusual firmness, drawing on both contract principles and constitutional due process. The State cannot take the benefit of a defendant's testimony and then disown its side of the bargain.

In Zani v. State, the Court of Criminal Appeals described immunity agreements as contracts in which each party gains a benefit and bears a burden, and it held that their terms must be strictly enforced and their inviolability “jealously guarded.” The federal due-process overlay comes from Santobello v. New York, which held that when a defendant's plea or cooperation rests in a significant degree on a prosecutor's promise, that promise must be fulfilled — even an inadvertent breach by a different lawyer in the same office does not excuse it, and the remedy may be specific performance.

But not every dismissal promise is an immunity grant. In Hatter v. State, the Court of Criminal Appeals clarified that although a grant of immunity is conceptually a promise to dismiss, it does not follow that every promise to dismiss is a grant of immunity. An immunity agreement carries a specific meaning, intertwined with the Fifth Amendment privilege and the Sixth Amendment rights it induces the witness to waive. A prosecutor's promise to dismiss one charge in exchange for a guilty plea to another is a plea bargain, not immunity — enforceable on its own terms, but analyzed under plea-agreement law rather than the immunity cases.

The line Hatter draws has a concrete payoff. Because an immunity agreement is defined by the exchange of testimony for protection, a defendant claiming immunity must be able to point to that bargain — a waiver of the privilege against self-incrimination in return for the State's promise — not merely to a prosecutor's statement that a case would be dropped. When the record shows a testimony-for-protection exchange approved through a dismissal, the immunity cases govern and the bar is strong. When it shows a charge-for-plea exchange, the plea-agreement cases govern, and the remedy for a breach is specific performance or plea withdrawal rather than dismissal on immunity grounds. Sorting the agreement into the right category is the first analytical step.

Immunized Testimony and the Independent-Source Problem

Use immunity does not end a case, but it reshapes what the State may do with it, and the mechanics matter to any witness who testifies under compulsion. The protection is only as strong as the government's ability to keep its later evidence clean.

Under Kastigar, once a person has testified under a grant of use and derivative-use immunity, a government that later prosecutes bears an affirmative burden: it must prove that the evidence it intends to use is derived from a legitimate source wholly independent of the compelled testimony. That is not a mere denial of taint; the prosecution must trace each step of its investigative chain and show it did not draw on the immunized words, either directly or as an investigative lead. The compelled testimony cannot be used as evidence, as a roadmap, or as a source of leads.

For a witness, the practical risk is real, because the information about what the government used lies largely in the government's hands. A witness who suspects that immunized testimony seeded the later case may struggle to prove it, which is why the burden rests on the prosecution and why a hearing on the independent source is the checkpoint. The defense response is to demand that the State document its pre-testimony evidence and its investigative sequence, so the court can test whether the case truly rests on independent proof. Where the State cannot make that showing, the derivative-use bar can exclude the tainted evidence even though the underlying prosecution was never itself forbidden.

The Procedural Vehicle and Burden of Proof

An immunity defense is designed to stop a prosecution before trial, so it is raised pretrial and litigated on a defined burden-shifting framework. The right vehicle is a motion to set aside or dismiss the charging instrument — the same family of motion this hub addresses.

In Zani, the court explained that because immunity agreements seek to avoid not just conviction but prosecution itself, they must be raised in a way that avoids the trial — a motion to dismiss the indictment based on the grant of immunity, decided at a pretrial hearing. The court set a two-step burden: the defendant must first show the existence of an immunity agreement by a preponderance of the evidence; once that is shown, immunity is treated like a defense, and the burden shifts to the State to show beyond a reasonable doubt why the agreement is invalid or why prosecution should be allowed despite it.

Two practical points follow. First, the hearing is the event that matters, so the defense should be prepared to prove the agreement, the defendant's performance, and the court approval that made it binding, all before trial; the State is generally confined to the evidence it presents at that pretrial hearing to defeat the agreement. Second, papering the deal well in advance — a written agreement, a clear record of the court's approval of the resulting dismissal, and proof of the testimony given — is what makes the bar enforceable later. An immunity promise that lives only in a hallway conversation is the kind Graham refused to enforce.

Plea, Non-Prosecution, and Other Prior Bars

Immunity is the sharpest of a family of “prior bar” defenses — grounds that defeat a prosecution because of something that happened, or was promised, before the current charge. Several deserve brief mention so counsel can route each to the correct analysis.

A negotiated plea or charge bargain can bar a later prosecution when the State agreed to dismiss or not refile a charge as part of the deal. Under Hatter and Santobello, such a promise is enforced as a plea agreement — requiring, in the Texas transactional context, the same court involvement that makes any dismissal binding — rather than as an immunity grant. A promise not to file charges before any case exists is treated as an exercise of prosecutorial discretion, and its enforceability turns on the presence of a genuine, provable agreement.

Two other prior bars are covered in companion satellites. A statutory discharge for the State's failure to provide a speedy trial can, in limited circumstances, bar further prosecution under article 28.061 — a bar sharply narrowed by later amendments, addressed in the article 28.061 satellite. And a prior jeopardy — an acquittal, a conviction, or a fact conclusively decided in the defendant's favor — bars re-prosecution under the Double Jeopardy Clause and the collateral-estoppel rule of Ashe v. Swenson, addressed in the double-jeopardy satellite. Each is a distinct doctrine, but all share this page's structure: a completed prior event, raised pretrial by a motion to set aside the charge, that the State cannot simply relitigate.

Routing each defense to its correct doctrine is not a technicality; it determines the standard, the burden, and the remedy. An immunity claim rises or falls on a court-approved agreement and the Zani burden-shift, a plea-agreement claim on Santobello enforcement, a double-jeopardy claim on attachment and sameness, and a speedy-indictment claim on the article 28.061 and article 32.01 framework. A motion that names the right bar and marshals the right proof is far more useful to the court than one that blends them together.

Where This Fits

This guide is one of four situations that arise under the Motion to Set Aside / Dismiss the Charging Instrument. 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.

• Kastigar v. United States, 406 U.S. 441 (1972) — Use and derivative-use immunity is coextensive with the Fifth Amendment; the government must later prove evidence came from a wholly independent source.
• Santobello v. New York, 404 U.S. 257 (1971) — When a plea rests significantly on a prosecutor's promise, the promise must be fulfilled; the remedy may be specific performance or plea withdrawal.
• Zani v. State, 701 S.W.2d 249 (Tex. Crim. App. 1985) — Immunity agreements are contracts raised pretrial by a motion to dismiss; the defendant proves the agreement, then the State must show its invalidity.
• Graham v. State, 994 S.W.2d 651 (Tex. Crim. App. 1999) — Texas immunity is statutory; transactional immunity requires court approval under article 32.02, and Texas recognizes no doctrine of equitable immunity.
• Smith v. State, 70 S.W.3d 848 (Tex. Crim. App. 2002) — A judge must approve the dismissal under an immunity agreement but need not know its specific terms; the nolle prosequi completes the agreement.
• Hatter v. State, 665 S.W.3d 584 (Tex. Crim. App. 2023) — Immunity authority derives from article 32.02, but not every promise to dismiss is immunity; a dismissal for a guilty plea is a plea bargain.

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