A statue of a mythological Chinese supernatural beast called Hsieh-Chai, who was believed to have the ability to tell the guilty from the innocent by butting them, stands at the University of Pennsylvania Law School.
“Slow and painful has been man’s progress from magic to law,” a proverb inscribed on the base says.
Three men convicted of murder by arson for a 1980 fire in Brooklyn are likely to be exonerated on Wednesday. One served 33 years. Another went blind in prison. The third died there.
Their convictions started with righteous rage over the death of a mother and her five children in a fire on Sackett Street. The New York of the 1970s and early 1980s was in the grip of arson fever, with more than 9,000 fires set annually, mostly in the Bronx and Brooklyn. (There were about 2,500 serious fires of any kind in 2014.)
What carried the three men into prison was not reliable evidence of an intentionally set blaze, but rather an arson investigation that was more like shamanism than science, rooted in hunches and folklore and disconnected from the dynamics of actual fires. Like the comparisons of bite marks, hair and handwriting, it was a forensic practice that had the authority of white-coat laboratory science but virtually none of its rigor.
“People didn’t understand the behavior of fire then,” said John J. Lentini, the author of “Scientific Protocols for Fire Investigation.” His December 2014 report on the Sackett Street fire said the original determination of arson was incorrect.
“They saw fire burning at one side of the apartment and at the other side of the apartment, and less burning in the middle,” Mr. Lentini said. “The interpretation was there were two separate fires. That’s just wrong.”
Since 1980, he said, fire scientists have come to understand the phenomenon of flashover, in which the gases from the initial point of a fire heat up a room until the entire space ignites. “That’s when you go from a fire in the room to a room on fire,” he said. Because different parts of a room might burn at varying intensities, fire investigators often mistakenly believed there had been two or more places where the fire began, a strong sign of arson.
The Sackett Street case was reopened when one of the three men, Amaury Villalobos, wrote in 2012 to Adele Bernhard, a lawyer and professor who runs alegal clinic at New York Law School that represents a small number of people with claims of innocence that cannot be proved by DNA testing alone. She asked Mr. Lentini to examine the original evidence and testimony. His conclusion — later ratified by experts consulted by the Brooklyn district attorney’s office — was that the Sackett Street fire investigation was plagued by systemic problems common to that era, making its findings useless.
“Much of what was believed by well-meaning investigators was, unfortunately, false,” Mr. Lentini wrote. “If today’s standards and knowledge of fire dynamics were applied to this investigation, the results would have been significantly different.”
That meant the pursuit of an innocence claim for Mr. Villalobos and his co-defendants, William Vasquez and Raymond Mora, would not be based on attacking the work of a single rogue detective or prosecutor but on the flaws in widely used investigative techniques.
Ms. Bernhard and her students presented their reinvestigation to a review panel set up by the Brooklyn district attorney, Ken Thompson.
“This is a very different sort of case for Thompson,” Ms. Bernhard said. “We’re not tracking down a police officer known to take shortcuts. This is where you’re saying the science has changed so much, they would never have been put on a case like that. That’s a big step.”
The introduction of DNA testing has shown that many earlier forensic techniques were shoddy facsimiles of actual science. Results could not be replicated; there were no acknowledged error rates; often the same people did both the investigations and the laboratory work, meaning that tests were skewed to fit conclusions.
In a suspected arson case in Georgia, two laboratories said there was no evidence of an accelerant. Nevertheless, the prosecutor, Nancy Grace, now a television commentator, won a murder conviction by introducing evidence that a trained dog had sniffed accelerants at the scene. In overturning the conviction, the appellate judges were reminded of a practice in ancient India involving a donkey put in a darkened room. The defendant had to touch its tail. If the beast brayed, that meant the defendant was guilty.
“It was the first lie detector test,” Mr. Lentini said.