On the morning of February 17, 2004, Cameron Todd Willingham was strapped to a gurney in the death chamber at the Huntsville Unit in Texas and executed by lethal injection. He was 36 years old. His last words, addressed to his ex-wife who had testified against him, were: "I am an innocent man, convicted of a crime I did not commit. I have been persecuted for twelve years for something I did not do." Then he was dead.

The State of Texas had convicted Willingham of capital murder — specifically, of deliberately setting a house fire on December 23, 1991, that killed his three daughters: Amber, two years old, and one-year-old twins Karmon and Kameron. Prosecutors told the jury that the fire science was conclusive: an experienced arson investigator had found clear evidence of accelerants, pour patterns, and "alligatoring" — the char patterns on wood that, experts claimed at the time, proved deliberate ignition. The jury deliberated for barely an hour. They sentenced him to death.

There was just one problem. The science was wrong. Not ambiguous — wrong. The "evidence" of arson that sent Willingham to death row was based on folklore and myth that the fire investigation community had systematically debunked throughout the 1990s. By the time Texas executed him, the nation's most respected fire scientists had concluded that the original investigation was riddled with errors. Two days before his execution, a report from a nationally recognized arson expert landed on Governor Rick Perry's desk. It concluded there was no credible evidence that the fire was arson at all.

Perry let the execution proceed. And then, when the Texas Forensic Science Commission began investigating whether junk science had put an innocent man to death, Perry replaced the commission's chairman — days before he was scheduled to hear damning testimony — to ensure the investigation would be neutralized.

This is not a story about an imperfect system doing its best. This is a story about a system that executed a man it had strong reason to believe was innocent, and then worked actively to make sure no one would ever know.

The Fire: What Actually Happened

December 23, 1991 was a cold morning in Corsicana, Texas, a small city about 60 miles southeast of Dallas. Cameron Todd Willingham, his wife Stacy, and their three young daughters lived in a cramped rental house on West 11th Avenue. That morning, Stacy had gone out early. According to Willingham, he was awakened by the screaming of his daughters and found the front part of the house engulfed in flames. He tried to reach the children's bedroom but was driven back by the heat. Neighbors saw him outside, distraught. He had burn marks on his hands and chest. The three girls — all under two years old — died in the fire.

Investigators from the Corsicana Fire Department and the Texas State Fire Marshal's Office arrived to examine the scene. What they found, they said, left no doubt: this was arson. Specifically, they identified what they described as classic arson indicators: a burn pattern that had supposedly moved in multiple directions from multiple points of origin; "alligatoring" of the wooden floor joists (a char pattern that resembles alligator scales); the presence of what they characterized as "pour patterns" suggesting accelerant had been splashed across the floor; a "crazed glass" pattern on the windows, which they said indicated rapid, intense heat from a deliberately set fire; and the fact that the fire had allegedly burned at an unusual low level along the floor rather than spreading from a single ignition point.

These indicators, the investigators concluded, were unmistakable signs that someone had poured an accelerant — gasoline or lighter fluid — and deliberately set the blaze. The only person with the means, motive, and opportunity, they decided, was Willingham himself. The motive they constructed: he was an abusive husband, a man with a troubled past who didn't really want to be a father. A jailhouse informant named Johnny Webb would later claim Willingham confessed to him. (Webb, it later emerged, had been promised a reduced sentence.)

What no one in the Corsicana courthouse told the jury was that every single "classic arson indicator" identified by the fire investigators had been debunked by peer-reviewed fire science research — in studies published years before Willingham's 1992 trial.

The Science That Wasn't: Arson Investigation's Dirty Secret

The field of arson investigation has, for much of its history, rested more on tradition and folklore than on empirical science. From the 1950s through the 1980s, fire investigators passed down a set of "indicators" that were said to distinguish accidental fires from deliberately set ones. These indicators were taught in training courses, codified in manuals, and repeated by expert witnesses in courtrooms across the country. They were also, in many cases, simply wrong.

Beginning in the 1970s and accelerating through the 1980s and 1990s, fire scientists — researchers who actually studied fire behavior under controlled conditions — began systematically testing these traditional indicators. What they found was alarming. The "pour patterns" that investigators attributed to deliberately splashed accelerants? Identical patterns could be — and regularly were — produced by accidental fires under the right conditions. "Alligatoring," supposedly proof of intense chemical accelerant? Caused by ordinary rapid fire growth. "Crazed glass"? Produced by ordinary thermal shock, nothing specific to arson. And the multiple points of origin that investigators said could only result from multiple deliberately set fires? Fire can burn in erratic, multi-directional patterns for entirely mundane reasons, including the way furniture, flooring materials, and air flow interact.

By 1992 — the year of Willingham's trial — the National Fire Protection Association had published NFPA 921, a guide to fire investigation that explicitly warned against relying on these traditional indicators as proof of arson. The document wasn't obscure; it was the definitive professional standard. But arson investigators across Texas, and across the country, largely ignored it. Their courtroom testimony continued to rely on the old myths. Juries heard it presented as settled science. Defendants went to prison — and in some cases, to death row.

This is the context in which Cameron Todd Willingham was convicted. The indicators that Manuel Vasquez of the Texas State Fire Marshal's Office identified as proof of arson were, by the published standards of the fire science community, not proof of anything. They were consistent with a fast-moving accidental fire — which is exactly what multiple scientists would later conclude the Willingham fire was.

"The Constitution of the United States is a law for rulers and people, equally in war and in peace, and covers with the shield of its protection all classes of men, at all times, and under all circumstances."

The Experts Speak — Too Late to Matter

By the early 2000s, Willingham had been on death row for over a decade, and his execution date was approaching. His court-appointed attorneys, working with minimal resources as is typical in capital cases in Texas, reached out to fire science experts. What they received back was extraordinary.

In early 2004, Dr. Gerald Hurst, a Cambridge-educated chemist and one of the country's foremost fire investigators — a man who had worked for both prosecution and defense in hundreds of cases and had been accepted as an expert witness in courts across the country — agreed to review the case file. He spent days going through the original investigation reports, photographs, and records. His conclusion was unambiguous: the evidence used to convict Willingham did not demonstrate arson. None of it. He identified specific, detailed errors in the original investigation. The "pour patterns" were not pour patterns. The "alligatoring" was consistent with an ordinary accidental fire. The "crazed glass" was the product of rapid thermal change, not chemical accelerants. The fire, he concluded, had most likely ignited in the children's bedroom — possibly from a space heater or another accidental ignition source — and spread naturally.

This report was submitted to the Texas Court of Criminal Appeals and to Governor Rick Perry's office. It arrived two days before Willingham's scheduled execution.

The Court of Criminal Appeals, which had the power to stay the execution, declined to act. Governor Perry, who had the power of clemency, declined to act. On February 17, 2004, Cameron Todd Willingham was executed.

The Innocence Project, which had also taken up the case, was stunned. They had expected at minimum a stay of execution to allow for review of the new evidence. Instead, the machinery of state death proceeded without pause.

After the Execution: The Commission and the Cover-Up

The story should have ended there — another execution, another closed case. But it didn't. David Grann, a staff writer at The New Yorker, spent years investigating the Willingham case, and in September 2009 published a landmark 15,000-word investigation titled "Trial by Fire." The piece meticulously documented the scientific failures in the original arson investigation, the suppression of the Hurst report, and the possibility — which multiple experts called a near-certainty — that the state had executed an innocent man.

The piece created a national firestorm. Texas had recently established the Texas Forensic Science Commission, an oversight body created in 2005 to review cases where forensic evidence was in question. The commission took up the Willingham case and commissioned its own independent review, which was assigned to Craig Beyler, one of the nation's top fire scientists and the technical director of Hughes Associates, a fire investigation firm. Beyler's report, delivered to the commission in August 2009, was even more damning than Hurst's. He concluded that the original arson investigation "did not comport with the scientific method and/or was not based on a thorough and reliable investigation." Every major arson indicator cited at trial was, Beyler found, either scientifically unsupported or demonstrably incorrect. The fire, he concluded, showed no reliable evidence of arson whatsoever.

Beyler was scheduled to present his findings to the full Texas Forensic Science Commission on October 2, 2009. It was expected to be a watershed moment — the first time an official Texas government body would formally hear that the state had likely executed an innocent man.

Four days before that hearing, Governor Rick Perry replaced the commission's chairman, Dr. Samuel Bassett, a prominent Austin attorney who had been presiding over the investigation. Perry appointed a new chairman: Williamson County District Attorney John Bradley, a hardline law-and-order prosecutor who had publicly opposed the use of post-conviction DNA testing in innocence cases and had delayed for years the release of a man later exonerated by DNA evidence.

Under Bradley's leadership, the commission immediately postponed the Beyler hearing. The investigation was effectively frozen. Bradley spent months questioning the commission's jurisdiction and working to narrow the scope of the review. When the commission finally did issue a report in 2011, it carefully avoided making any direct finding about whether Willingham was innocent or whether the original arson investigation was faulty — despite the overwhelming scientific consensus that it was both.

Rick Perry's "Monster" and the Politics of Innocence

When David Grann's article and the subsequent commission investigation generated sustained national press attention, Governor Perry did not respond with reflection. He responded with contempt. In October 2009, Perry publicly called Cameron Todd Willingham "a monster." He accused Willingham's advocates of using "supposed experts" to "muddy the waters." He defended the execution as just and correct, and he questioned the motives of everyone who disagreed.

Perry's response is worth examining closely, because it illustrates exactly how the political incentive structure of American criminal justice works against accountability. For Perry, admitting error — or even acknowledging uncertainty — carried catastrophic political risk. He was already considering a presidential run (which materialized in 2012 and again in 2016). To concede that Texas might have executed an innocent man would be to hand his opponents a devastating weapon. The alternative — to brand Willingham a "monster" and the scientists questioning the conviction as partisan advocates — carried no political cost at all. In Texas's political environment, defending an execution against soft-hearted scientists was a feature, not a bug.

This is the mechanism by which states resist post-execution accountability. It is not always, or even primarily, driven by bad faith about the facts. It is driven by the rational calculation that politicians who admit fallibility in capital cases lose elections, while those who double down win them. The system provides no incentive for honesty and every incentive for denial.

Perry would go on to replace Dr. Bassett with John Bradley — who, before being appointed TFSC chairman, had prevented a man named Michael Morton from being tested for DNA evidence for years. Morton was exonerated in 2011 after spending nearly 25 years in prison for a murder he did not commit. The prosecutor who had withheld evidence in Morton's case was later found by a court of inquiry to have committed criminal contempt. Bradley, Perry's chosen champion for the Willingham investigation, had been an active obstructionist in that case.

The circle of accountability in Texas's justice system has always been very small, and the people standing inside it have always been very well protected.

The Jailhouse Informant: Johnny Webb's Confession

The prosecution's case against Willingham rested on two pillars: the arson evidence and the testimony of a jailhouse informant named Johnny Webb, who claimed Willingham had confessed to him while they were both incarcerated in the Navarro County jail awaiting trial. Webb testified that Willingham told him he had set the fire to collect insurance money — though investigators had noted that Willingham actually carried very little insurance on the property.

In 2014, a decade after Willingham's execution, Webb recanted his testimony in a sworn affidavit. He said he had fabricated the confession story. He said he had received a favorable deal on his own case in exchange for his cooperation — including a sentence reduction and a transfer to a more comfortable facility. He said he had been coached on what to say. He said he had lied, and that Cameron Todd Willingham was innocent.

By that point, it didn't matter. The man he had helped send to death row had been dead for ten years.

Jailhouse informants are one of the most thoroughly documented sources of wrongful convictions in the American legal system. Studies by the National Registry of Exonerations and the Innocence Project have consistently found that informant testimony — typically provided by incarcerated individuals seeking sentence reductions — appears in a disproportionate share of wrongful conviction cases. The incentive structure is self-explanatory: the informant has everything to gain by providing useful testimony and nothing to lose if it turns out to be false. In most cases, there is no mechanism for verifying the claim. And in capital cases, where prosecutors are under intense pressure to secure convictions, the temptation to rely on informant testimony — even when it is unverifiable or suspicious — is particularly strong.

Navarro County prosecutors knew Webb had received benefits in exchange for his testimony. Whether they adequately disclosed this to the defense, as required by Brady v. Maryland, 373 U.S. 83 (1963), has never been fully adjudicated. Under Brady, the prosecution is constitutionally required to disclose material exculpatory evidence — including the existence of benefits given to government witnesses. Failure to do so is a Brady violation, which can form the basis for overturning a conviction. But Cameron Todd Willingham is dead. There is no conviction to overturn.

"The administration of justice is the firmest pillar of government."

What the Law Allows — and What It Doesn't

The Willingham case exposes a series of legal doctrines and institutional structures that make it extraordinarily difficult for courts to correct errors — and nearly impossible to correct them posthumously.

The finality doctrine. American courts have a deep institutional commitment to the finality of criminal judgments. Once a conviction has been affirmed on direct appeal and habeas corpus proceedings have been exhausted, the legal system treats the matter as settled. This commitment to finality serves legitimate purposes — the justice system cannot function if every case remains permanently reopened — but it exacts a terrible cost when the original judgment was based on false evidence. The Supreme Court addressed this tension in Herrera v. Collins, 506 U.S. 390 (1993), where the majority held that a claim of actual innocence, standing alone, is not grounds for federal habeas corpus relief absent an independent constitutional violation. In other words: if the procedures were followed correctly, the fact that you are innocent may not be enough to get federal courts to intervene.

Forensic science standards in court. The standard for admitting expert testimony in federal courts — and in most state courts — is governed by Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), which requires that expert testimony be based on sufficient facts, reliable methods, and proper application of those methods to the case. But Daubert's gatekeeping function depends on trial judges who understand science well enough to identify unreliable methodology — and in the early 1990s, when Willingham was tried, few judges were equipped to scrutinize the traditional arson investigation methodology. The folklore of fire investigation had been admitted in court so many times, by so many witnesses, that it had acquired a presumption of reliability it had never earned.

Executive clemency. The governor's power of clemency — commutation or pardon — exists precisely as a safety valve for cases where the legal system has failed. In Texas, the governor can only grant clemency if the Board of Pardons and Paroles recommends it. The board did not recommend clemency for Willingham. Whether it even reviewed the Hurst report before his execution remains unclear. Perry did not request such a review.

Absolute prosecutorial immunity. Under Imbler v. Pachtman, 424 U.S. 409 (1976), prosecutors have absolute immunity from civil liability for actions taken in the course of their prosecutorial duties — including, courts have held, the knowing use of false testimony and the suppression of exculpatory evidence. This doctrine means that even if a prosecutor committed flagrant Brady violations in the Willingham case, there would be no civil remedy available to his family. The only accountability mechanism would be bar discipline — which is rarely pursued, and rarely results in meaningful sanctions even when it is.

How Many Others?

The Willingham case is not an isolated anomaly. It is the visible tip of a much larger iceberg. Research by the Innocence Project has identified multiple other cases in which convictions rested on arson evidence that has since been scientifically discredited. The Brandon Dassey and Steven Avery cases in Wisconsin, while not arson cases, illustrate similar dynamics: convictions built on informant testimony, coerced confessions, and questionable forensic science. The forensic disciplines with the least scientific foundation — hair comparison, bite mark evidence, fire investigation — appear repeatedly in wrongful conviction case files.

A 2009 report by the National Academy of Sciences, titled "Strengthening Forensic Science in the United States: A Path Forward," concluded that many forensic disciplines widely used in courts — including arson investigation — lacked adequate scientific validation. The report called for urgent reform, including mandatory accreditation of forensic labs, uniform standards, and independent oversight of forensic scientists. Congress received the report. It largely ignored it.

Texas, in particular, has a documented history of capital cases with troubling forensic questions. The state has executed more prisoners than any other in the modern era of capital punishment. The combination of a prosecutorial culture focused on obtaining convictions, a courts system with limited resources for post-conviction review, and a political culture that treats any scrutiny of executions as weakness creates conditions that are, as Willingham's case demonstrates, capable of producing irreversible injustice.

The number of people executed in the United States on the basis of false or misleading forensic science will never be fully known. Once someone is executed, the investigation largely stops. There is no ongoing review process for closed capital cases. The Texas Forensic Science Commission, even in its most aggressive phase, was reviewing one case at a time, in response to media pressure, without any systematic process for identifying other cases where similar problems might exist.

"The true administration of justice is the firmest pillar of good government... the laws have nothing to expect but the contempt of those who are governed by them."

Stacy Willingham's Silence — and What It Means

One of the more painful dimensions of this story involves Stacy Willingham, Cameron's ex-wife, who divorced him during his time on death row and who testified at trial in a manner that was damaging to him — suggesting, among other things, that he had shown insufficient grief after the fire and had been focused on moving a dartboard out of the house as their daughters burned. (Willingham maintained he was attempting to break a window to reach the children; investigators later said the evidence was consistent with this account.)

After Grann's article brought renewed national attention to the case, Stacy spoke publicly, and her perspective was complex and shattering. She said she had been manipulated by investigators who pressured her to characterize her ex-husband's behavior as suspicious. She said she had not been shown or explained the exculpatory fire science. She said she did not know, when she testified, what she was being asked to help accomplish. And she said — this is perhaps the most devastating detail — that she had believed, at various points, that Todd might actually be innocent.

Stacy Willingham was herself a victim of the Corsicana fire. She lost her three daughters. And then the system that was supposed to give her justice used her grief and her testimony to execute a man who may not have been responsible for their deaths. She was never given the real facts. She was a means to an end.

This is what the justice system does to victims when it is more invested in convictions than in truth. It appropriates their suffering. It weaponizes their grief. And it leaves them, when the truth eventually comes out, with something that cannot be undone: the knowledge that they helped the state kill someone who may have been innocent.

The National Reckoning That Never Came

In the years following the Grann article and the TFSC investigation, many observers expected a serious national conversation about the reliability of arson evidence and the safety of capital punishment. That conversation happened in academic circles, in policy organizations, and in the pages of quality journalism. It did not happen in legislatures or in the courts to any meaningful degree.

The National Fire Protection Association updated NFPA 921 repeatedly in the years following Willingham's execution, incorporating additional scientific research and further discrediting the indicators used at his trial. The document is more clearly and authoritatively written today than it was in 2004. It has not triggered any systematic review of arson convictions in Texas or most other states.

The FBI, in 2015, released a landmark review finding that hair comparison testimony — another forensic discipline with questionable scientific foundations — had been flawed in a majority of the cases in which FBI agents had testified in the preceding decades. The review covered hundreds of cases, including some capital cases. The agency notified defendants and their attorneys. Follow-through was inconsistent; many notifications went unheeded.

No comparable review has been conducted for arson convictions. The Texas Department of Criminal Justice does not maintain a public database of cases where arson evidence has been subsequently questioned. The Texas Court of Criminal Appeals, which has the power to order such reviews, has not done so systematically. The legislature has not funded it. And the political class — which would have to vote to fund and mandate such a review — has no incentive to do so, because the political cost of acknowledging the possibility that the state killed innocent people is still, in Texas in 2026, prohibitively high.

The Posthumous Exoneration That Never Came

In some states, the legal system has developed mechanisms for posthumous exoneration — official findings that a deceased person was wrongly convicted. These are rare. They are procedurally complex. They carry no legal remedy — the person is dead — but they carry enormous symbolic and historical weight. They are also, in many states, politically toxic, which is why they so rarely happen.

Texas has no clear legal mechanism for posthumous exoneration. The Board of Pardons and Paroles can theoretically recommend a posthumous pardon, and the governor can grant one, but neither has been willing to do so in the Willingham case. The Texas Court of Criminal Appeals has no clear authority to issue a posthumous finding of actual innocence. And without such a finding, the official record of the State of Texas will continue to reflect that Cameron Todd Willingham was a man who murdered his three daughters.

His family — his parents, his sisters — have spent more than two decades trying to clear his name. They have spoken to reporters, testified before commissions, and watched as each avenue for official acknowledgment closed. As of this writing, Cameron Todd Willingham remains, in the eyes of the State of Texas, a convicted murderer who received a just punishment.

This is the final cruelty of a system without accountability: the official record persists even when it is demonstrably false.

A Reform Blueprint: How to Prevent the Next Cameron Todd Willingham

The Willingham case is not primarily a story about bad people, though there were surely failures of character at multiple points. It is a story about systemic structures that produce unjust outcomes reliably and predictably, and that resist correction because the people in positions of power have no incentive to accept the costs of accountability. Reform requires changing those structures.

1. Mandatory scientific validation for forensic disciplines before courtroom use. Congress should establish, through statute, a requirement that any forensic methodology used in criminal trials must meet minimum standards of scientific validation established by an independent body with no connection to law enforcement or prosecution. The National Institute of Standards and Technology (NIST), which has taken on some forensic science validation work, should be empowered and funded to complete that task systematically, including for arson investigation, hair comparison, bite mark evidence, and other disciplines with documented reliability problems.
2. Retroactive review of convictions based on since-discredited forensic science. Congress and state legislatures should mandate and fund a systematic review of all capital and life-sentence convictions in which arson investigation, hair comparison, or other identified unreliable forensic methodologies were used. This review should be conducted by independent forensic scientists, not law enforcement agencies, and defendants — or their families, in the case of posthumous cases — should be notified of findings and given access to counsel for potential post-conviction proceedings.
3. Strict regulation of jailhouse informant testimony. State legislatures should enact statutes requiring pre-trial disclosure of all benefits, promises, and prior testimony given by any jailhouse informant. Courts should be required to conduct reliability hearings before such testimony is admitted in capital cases. Prosecutors who fail to disclose informant benefits should face mandatory bar discipline, not merely discretionary review. Multiple states — including Illinois and California — have enacted partial reforms in this area; they should be adopted nationwide and strengthened.
4. Gubernatorial clemency review reform. The executive clemency process in capital cases should include a mandatory scientific review component whenever a defendant presents credible expert evidence of forensic error. A governor should not be permitted to proceed with an execution after receiving a credible scientific challenge to the underlying forensic evidence without a documented, public response to that challenge. The failure to engage with the Hurst report before Willingham's execution was not an accident of procedure; it was an exercise of political will. Procedure should be designed to make that choice impossible.
5. Firewall between forensic science oversight commissions and the executive branch. Texas Forensic Science Commission members should serve fixed terms and be removable only for cause, not at the governor's pleasure. The events of September 2009 — in which a governor replaced a commission chairman four days before politically inconvenient testimony — should be structurally impossible. Commission chairs should be selected by the commission itself, not appointed by the governor.
6. Formal posthumous exoneration process in every state. Every state should have a clear legal pathway for seeking a posthumous finding of actual innocence, with standing granted to family members of deceased individuals. Such findings should trigger an obligation on the state to correct official records, publish a formal acknowledgment, and provide reasonable compensation to immediate family members. The current situation — in which a man executed on the basis of false science remains officially guilty forever — is not consistent with constitutional values or basic human decency.
7. Abolish or severely restrict absolute prosecutorial immunity in cases of capital misconduct. The doctrine of absolute prosecutorial immunity, as established in Imbler v. Pachtman, should be revisited by Congress and the courts in the context of capital cases. Where a prosecutor knowingly used false testimony, suppressed Brady material, or committed other misconduct in a case that resulted in a death sentence, the shield of absolute immunity should not apply. Civil liability for the most egregious misconduct is a necessary component of meaningful deterrence.
8. Require open-file discovery in all capital cases. Prosecutors in capital cases should be required by law to provide defense counsel with complete, unrestricted access to all investigative files — not just materials the prosecution decides are material and favorable to the defense. The current Brady disclosure standard, which depends on prosecutorial judgment about what is "material," has proven demonstrably inadequate in case after case. Open-file discovery removes that discretion entirely.
9. Establish a National Innocence Commission. Modeled on the National Transportation Safety Board, a National Innocence Commission should be established with authority and funding to conduct independent, systematic reviews of potential wrongful conviction cases — including capital cases, including posthumous cases. The commission should have subpoena power, independent legal staff, and a mandate to identify not just individual failures but systemic patterns that produce unjust outcomes.
10. Moratorium on executions pending forensic review. No state should execute an individual whose conviction was based in whole or in substantial part on forensic evidence that has been subsequently questioned by credible scientists, unless and until an independent scientific review has confirmed the reliability of that evidence. The execution of Cameron Todd Willingham — carried out over the explicit objections of one of the country's top fire scientists — should be the last time this happens in America.

Conclusion: The State Owes More Than Silence

Cameron Todd Willingham was executed twenty-two years ago. His daughters are still dead. The question of what actually killed them — accident or arson — is, in the view of every qualified fire scientist who has reviewed the case, no longer seriously in doubt. The fire was not arson. The evidence was folklore. The conviction was a mistake. And the mistake was compounded, deliberately and cynically, by a governor who valued his political future more than the truth, and by a legal system that has no mechanism for confronting its worst errors.

His parents, Eugenia and Jesse Willingham, spent years fighting for their son's name. His sisters have spoken to reporters across the country. None of them have received so much as an acknowledgment from the State of Texas that it is possible — possible, let alone probable — that the man it killed was innocent.

This is what impunity looks like. It is not always dramatic. It is not always a conspiracy. Sometimes it is just a system — of incentives, of immunities, of finality doctrines and political calculations — that makes it easier to maintain a lie than to confront a truth. And in that quiet, bureaucratic way, it is perhaps more frightening than any individual act of corruption.

The Founders understood that the justice system, left unaccountable, would become an instrument of tyranny. They built a Constitution designed to prevent that outcome. They could not have imagined a governor replacing a forensic science commissioner four days before inconvenient testimony, or a Supreme Court holding that actual innocence might not be enough to override the finality of a death sentence. But they knew the impulses — toward self-protection, toward power, toward the suppression of inconvenient truth — and they knew those impulses would require constant resistance.

Cameron Todd Willingham is dead. The three little girls are dead. The jailhouse informant who lied about them has confessed. The fire experts have spoken. The only thing left is for the State of Texas — and for the country — to decide whether accountability is a principle we actually believe in, or just a word we use when it costs us nothing.