John Lentini’s new book is ostensibly about arson cases. In truth, it is about something far larger: the habits of mind by which criminal justice systems turn uncertainty into guilt.

John Lentini’s new book, Twice Burned: A Fire Scientist’s Adventures in Overturning and Preventing Wrongful Arson Convictions, is far more than a specialist’s account of fire investigation. It is an important book about professional error, cognitive bias, institutional overconfidence, and the terrible human cost of wrongful conviction.

Lentini is no marginal technician in a narrow field. He is one of the central figures in the modern development of fire investigation in the United States. Over decades, he worked in laboratories, at fire scenes, in courtrooms, and within the bodies that shaped professional standards. He helped challenge myths long treated as settled knowledge, exposed the role of bad science in wrongful convictions, and pushed his field toward something both more modest and more rigorous: the scientific method.

That is precisely what gives the book its broader significance. Lentini does not merely narrate dramatic cases. He exposes the logic that produces failure. He shows how investigators become attached too quickly to a theory; how inconvenient facts are pushed aside; how hypotheses harden into certainties; and how entire systems gradually join the same mistake. Though the book’s raw material is American arson litigation, its lesson extends far beyond arson. It is a book about law, evidence, humility, and the danger of false certainty.

One of the most important concepts associated with Lentini’s work is negative corpus. The idea is simple, and devastating. Instead of proving what happened, the investigator concludes that the incriminating explanation must be correct because no satisfactory alternative explanation has been found. Guilt is not established by evidence; it is inferred from the absence of another story.

This must be distinguished from careful reasoning by elimination. Serious investigation often does exclude possibilities. But negative corpus is not disciplined exclusion within a broader evidentiary structure. It is the point at which the absence of an alternative ceases to signal the need for more inquiry and begins to function as evidence against the suspect. Instead of saying, “We do not yet know enough,” the system says, “If we have no other explanation, then this must be the one.” That is a dangerous move from evidentiary absence to positive conclusion.

At first glance, this may seem like a problem peculiar to fire scenes. It is not. It is a broader pattern of criminal reasoning, one that recurs wherever suspicion outruns proof. That is why Lentini’s book matters far beyond the world of arson. It gives us an exact vocabulary for a failure of thought that appears in homicide cases, sexual offense prosecutions, false confessions, and any criminal proceeding in which an institution finds uncertainty difficult to tolerate.

This is where the question of safety in criminal justice becomes crucial. In aviation, medicine, transport, and industry, it is long understood that error is not merely a personal failure. It is also the product of poorly designed systems. That is why those fields developed checklists, redundancies, debriefings, bias-reduction methods, and safety protocols meant to catch the mistake before it becomes catastrophe.

Criminal justice, by contrast, too often treats error as exceptional and accidental, rather than as a foreseeable consequence of inadequate safeguards. If advanced safety methods were taken as seriously in criminal law as they are in aviation or surgery, many notorious cases would likely have unfolded very differently. Some might never have become indictments at all.

The point becomes clearer when one turns to interrogation. In the 1990s, England recognized that pressure-based interrogation methods were no more likely to produce true confessions than false ones. In response, it developed the PEACE model, built on open-ended questioning, evidence disclosure, and critical testing of the suspect’s account without threats, promises, or intimidation. Its premise is both simple and civilized: the goal of interrogation is not to break a suspect, but to find the truth while minimizing the risk of false confession.

Against that stands a very different mentality, one that effectively begs the question from the outset: if a suspect is being interrogated, he is probably guilty; if he is probably guilty, pressure is justified; and once a confession is obtained, the inquiry is effectively over. In such a system, the confession is no longer one item within a body of evidence to be critically examined. It becomes the event that ends criticism.

That is exactly why a concept like negative corpus matters so much. It captures a deeper pathology. The danger is not only that we do not know what happened. The danger is that a poorly designed system cannot bear not knowing. It fills the gap with guilt.

A particularly sharp illustration of this pattern appears in the case of Nissim Haddad. The central question there was whether a certain medical finding compelled a criminal conclusion, or whether the absence of a clear alternative explanation was still insufficient to convict. Justice Yitzhak Amit wrote: “Even if this does not in itself constitute evidence sufficient to convict the appellant, …, [still] the absence of a reasonable alternative scenario for the tears caused in the infant’s anus leads to one possible conclusion, namely the incriminating conclusion.”

Set against this was the response of Dr. Chen Kugel, then the senior pathologist whose opinion shook the case, and now Director of the National Institute of Forensic Medicine: “Indeed, it is unclear what did happen to the baby. Accordingly, the investigative authorities should be so kind as to investigate what did happen. The fact that matters are unclear is certainly not a reason to place an innocent person in prison.”

These two statements embody two rival conceptions of criminal adjudication. According to the first, the absence of a satisfactory alternative strengthens conviction. According to the second, the absence of clarity is a reason for more investigation, not punishment. The first turns uncertainty into adjudication. The second refuses to perform that conversion. That is why the Haddad case is such a clean example of negative corpus: not because guilt was proven, but because the lack of an alternative began to do the work of proof.

If Haddad illustrates that pattern clearly, the Zadorov case illustrates an even graver and more complex one. Beyond the immense public controversy, it reveals what happens when several pathologies converge: tunnel vision, the narrowing of inquiry around a single theory; the shrinking willingness to examine potentially exculpatory evidence; and the presentation of dependent pieces of evidence as though they were independent corroboration.

Consider the issue of the work pants. Police searched the landfill for many days and collected several pairs of pants. But when it became clear that the pants found did not bear the victim’s blood, an order was given to destroy them—without first determining which of them belonged to Zadorov, and without allowing him to rely on the straightforward possible inference that there was no connection between his pants and the murder. This was not merely an evidentiary mishap. It was a revealing moment in which a one-directional investigative theory narrowed, in advance, the willingness to seriously examine a potentially exculpatory fact.

A fair investigation does not fear exculpatory evidence. It seeks it. An investigation suffering from tunnel vision does the opposite: everything that strengthens the theory receives attention; everything that weakens it is pushed aside, reinterpreted, or deprived of weight. At that point, even the absence of incriminating evidence no longer works in the suspect’s favor. The system becomes more loyal to its theory than to reality.

But there is something still worse than negative corpus: circular reasoning designed to prove what was assumed from the outset. The prior assumption is guilt. From there, every fact is interpreted accordingly. If incriminating evidence exists, it proves guilt. If such evidence does not exist, that too proves guilt, because the suspect must be more sophisticated—he removed the traces, cleaned up after himself, outsmarted the investigators.

This is where inquiry becomes nearly unfalsifiable. If there is evidence, he is guilty. If there is no evidence, he is a clever guilty man. If he breaks under interrogation, he is guilty. If he withstands pressure, he is calculating. If no footprints are found, he removed them. If there is contradiction, it shows guilt. If there is no contradiction, it shows preparation. At that point, the investigation is no longer testing a hypothesis. It is protecting one.

The absence of Zadorov’s shoeprints at the scene illustrates this well. A cautious reading would treat that absence as a serious difficulty for the incriminating theory, or at least as a reason for evidentiary modesty. But once one infers instead that he must have removed the prints, the absence itself becomes incriminating. The system is no longer measuring the theory against the facts. It is immunizing the theory against disproof.

This connects to another deep problem: reliance on one piece of evidence and its offspring as though they were several independent pieces of evidence. In the appeal judgment affirming Roman Zadorov’s conviction, Justice Amit wrote: “The confession to the cellmate stands on its own. The confession to the investigators stands on its own. The reenactment too stands on its own. Before us are three separate and independent layers.”

But that is precisely the difficulty. Those three “layers” do not come from three independent sources. They come from the same person, within a single interrogative sequence. They are therefore not truly independent corroborations, but different manifestations of the same source, subject to the same pressures, influences, and possibilities of error or false confession.

Cesare Beccaria saw the problem clearly more than two and a half centuries ago: “When the proofs of a fact depend upon each other, that is, when the evidence is not capable of being verified except by one another, then the more proofs are adduced in support of it, the less is the probability of the fact; because whatever causes the first proofs to fail must also cause the subsequent proofs to fail.”

Not every multiplication of evidence is genuine strengthening. When several pieces of “evidence” ultimately rest on the same foundation, they do not necessarily add force. They may simply replicate the same weakness in several forms.

Sometimes the system goes further still. It not only treats dependent evidence as independent. It also finds itself unable to tolerate even the suggestion of innocence. I learned this personally. Three weeks after my arrest in March 2003, during the hearing on my detention until the end of proceedings, the prosecutor told the judge that she “took her hat off” to me for having withstood all the interrogation tricks. My mother burst out with the simplest explanation of all: “Of course he withstood the tricks. Because he is telling the truth.” Instead of reconsidering the incriminating assumption, the system silenced the voice that offered the most direct and human explanation. The judge did not suggest that the prosecutor might consider that possibility. He threatened to remove my mother from the courtroom. In that small moment, a larger pathology was laid bare: the system often does not merely struggle to see an exculpatory possibility—it struggles even to bear hearing it spoken aloud.

That, ultimately, is the larger lesson of John Lentini’s book. Its importance lies not only in what it teaches about fires. It lies in the language it gives us for recognizing dangerous intellectual and institutional processes far beyond arson: negative corpus, tunnel vision, circular reasoning, excessive confidence in expertise, tendentious interpretation of facts, and the replacement of proof with conjecture. The real danger is not merely isolated error. It is a style of thinking that cannot tolerate doubt and rushes to close the case with guilt.

That is why Twice Burned matters so much to readers far from the world of fire science. It is a reminder that safety in criminal justice is not a metaphor. It is a basic condition of liberty. Just as we would never entrust an aircraft, an operating room, or an industrial plant to a system not designed to reduce error, so too we should not accept a criminal process that is not designed to reduce the risk of wrongful conviction. In that sense, Lentini’s book is not only an important professional work. It is a moral warning.

© The Times of Israel (Blogs)