Arson Convictions, Fire Investigations Feel the Heat
As decades of flawed and unscientific fire investigation techniques call arson convictions into question, new recipes emerge for a system-wide overhaul.
Phoenix attorney Larry Hammond knows just how much is riding on the paperwork on his desk. The chairman of the Arizona Justice Project, he is fighting hard to overturn the conviction of Louis C. Taylor, imprisoned 38 years ago for intentionally setting a catastrophic fire at Tucson's landmark Pioneer Hotel that ultimately killed 29 people. Ever since Hammond first studied the case in 1999, he's been convinced that the blaze was not arson.
Taylor, now in his 50s but then a 16-year-old petty thief known to hang around pool halls, claimed he slipped into the hotel that December night to cadge free food and drinks in the ballroom. After the fire broke out, he was spotted nearby with matches in his pocket.
Unfortunately for Taylor, 1970 was the dark ages of fire investigation. Arson investigator Cyrillis W. Holmes Jr. found no tangible evidence of how the fire started, yet he divined from burn patterns and fire debris that at least two fires had been deliberately set about 60 feet apart in the fourth-floor hallways.
Multiple points of origin are a powerful indicator of arson — if they're real. But old-time fire investigation was a mix of old wives' tales, myths and oral hand-me-down wisdom with no science behind it. Fire debris was read like tea leaves. And marks on floors and carpets called "pour patterns" were routinely interpreted as points of origin and cited as evidence of an arsonist pouring liquid accelerants like gasoline.
Since then, science has rendered such "arson indicators" obsolete and convictions based upon them questionable.
It is one thing to suspect that fallacious evidence helped convict someone like Taylor, quite another to secure a new trial or exoneration because of "new" scientific evidence. But Larry Hammond isn't cowed by a tough fight, and buoyed by a "breathtaking" report from renowned independent fire scientist John Lentini, he's readying a petition to file in state court.
He knows he has a tough road ahead.
"Our legal system is designed to foreclose post-conviction review," he noted, "and it does frustrate, and has frustrated, many of these cases." Attorney Walter Reaves, who works with the Innocence Project of Texas, agreed: "You have to convince a court that it is (newly discovered evidence), and then you have to convince them to actually listen and hear it."
DNA is the undisputed gold standard for exonerations, a virtually unassailable magic bullet. But arson convictions are a new frontier for exoneration work, and they are qualitatively different. If you find a bullet or knife in a dead man's back, no one disputes that a crime has taken place. Fires, however, are not so simple.
In America, while suspects are presumed innocent until proven guilty, frequently fires are not presumed accidental until proved to be arson. All fires necessitate an extra investigative step — an independent, science-based determination of arson to first ascertain that a crime even took place.
Civil cases often bifurcate issues of liability and damages, making the jury's first task to determine if a fire was intentionally set. Criminal arson cases seem to cry out for the same approach, given that a determination of arson alone can never definitively tie a person to a crime scene or unequivocally reveal a perpetrator's identity.
Yet any serious proposal for a similar system would be met, Lentini surmised, with "universal screaming and gnashing of teeth," and allegations that criminals were being allowed to slip through the net.
Gnashing teeth notwithstanding, the need for reform is critical. Lentini cites figures of 75,000 suspicious fires every year — "That's 75,000 chances to get it wrong," as he told CNN's Anderson Cooper.
According to a 2002 Bureau of Justice Statistics report, in just half the states in the U.S., more than 5,000 people are in prison for arson crimes. Arson is the only crime for which someone can receive the death penalty based on the testimony of an expert witness whose education ended with high school. And although no one spoken to for this article would hazard a specific estimate of how many innocents are imprisoned on arson convictions, answers ranged from "dozens to hundreds" to "tons."
If ever a case embodied the disastrous consequences of the obsolete beliefs about fires, it is Cameron Todd Willingham's. The Texan was convicted in 1991 of the arson murders of his three children, all under age 3. In 2004, Willingham's appellate lawyer — Walter Reaves — commissioned a review by chemist Gerald Hurst, a key player in the fire investigation wars. Hurst's powerful report debunked all 20 so-called "arson indicators" used to convict Willingham. He was executed anyway.
While it's too late for Willingham, the New York Innocence Project was galvanized to create an Arson Review Committee corralling five top experts, including Lentini and veteran investigator Douglas Carpenter, to compare Willingham's case to that of Ernest Willis, a fellow Texan convicted on similar evidence of setting a fire that killed two sleeping women in a house he was staying in. After 17 years in prison and only months after Willingham's execution, Willis was exonerated and released.
The Arson Review Committee's 2006 report echoed Hurst in discrediting all the "arson indicators" found by deputy state fire marshal Manuel Vasquez at the Willingham fire. Patterns Vasquez attributed to ignitable liquid or accelerant, it concluded, could not be used to distinguish arson from an accidental fire.
The report is now under review by the Texas Forensic Science Commission, whose chair, Sam Bassett, told Miller-McCune.com that it has voted to hire an expert to lead them in their investigation. The commission is charged by statute with conducting investigations. If they concur with all the other experts, they could make recommendations for further review or even for system reform.
"I believe it would be within our purview to comment upon any broader issues such as the possibility of misconduct or professional negligence in other cases," Bassett wrote in an e-mail. "However, that is dependent first upon our finding that misconduct or professional negligence occurred in the Willis/Willingham case. Until we receive feedback from our expert, the Commission will remain neutral on this issue and there will be no further comment until such time as we issue our report."
The possibility of misconduct and professional negligence?
"Boy, there's a political football for you," Hurst said skeptically, imagining the ramifications of an official admission that an innocent man was executed.
Carpenter is more optimistic. "I'd certainly hope something tangible comes from the process," he said. "I think they're taking it seriously."
Yet Reaves, the attorney, worries that the commission's very review could do more harm than good: "I just have a hard time envisioning a state sponsored commission coming down and saying, 'Oops, we killed a guy that we shouldn't have killed." He predicted that it will conclude — wrongly — that Willingham was an isolated incident, or a problem that has been remedied and won't happen again.
Bringing in Real Science
Larry Hammond preferred instead to pin his optimism to an imminent National Academy of Sciences report on forensic science reform. He expects it to "trash some sacred cows of forensic science," with arson at the top of the list. Perhaps, the most hopeful and immediately pragmatic agent of change, however, is the Arson Screening Project launched in July 2008 at The Center for Modern Forensic Practice at John Jay College in New York, in consultation with the New York Innocence Project.
It will step into uncharted territory by assessing the scope of arson convictions based on accidental fires. It will also triage meritorious cases for review by the affiliated Arson Review Group of fire scientists, led by Lentini, and eventually will suggest a reform plan. Its director James Doyle said he will use his precious experts judiciously, weeding out time-waster cases where, say, a debunked "arson indicator" meets "a guy found on the crime scene with his eyebrows burned off and a gasoline can in his hand."
Wrongful convictions are only half the story. Only a multifaceted approach can overhaul a system where, for example, so many fire investigators making judgments of arson have little or no scientific training, and where discredited arson evidence is still presented in courts.
Science officially staked its claim in fire investigation in 1992 with the ground-breaking, peer-reviewed National Fire Protection Association code NFPA 921, the industry standard. By then, a cadre of forward-thinkers like Lentini, Hurst, Carpenter and John DeHaan, author of the Kirk's Fire Investigation handbook, doggedly had been using science with full-scale recreations, test burns and laboratory experiments to study fire's behavior. Incredibly, until then, no one had actually compared the aftermaths of accidental and intentionally set fires.
Industry acceptance of NFPA 921 proved glacially slow. The emerging scientific information was rejected, disputed and dismissed. Tensions ran high between old guard fire investigators and the science guys. DeHaan can recall receiving hate mail. Old-school investigators weren't bad people, though, just working with bad information. And some defensiveness about potentially having had a hand in wrongful convictions was certainly understandable. Less easy to understand: those diehards still resisting science-based knowledge 16 years later.
In 2006, investigator Cyrillis Holmes told the Arizona Daily Star that he stands by his original opinions about the Pioneer Hotel fire. But insurance investigator Marshall Smyth, who also worked on the case, has since likened himself and Holmes back then to members of a black magic society. "If that fire were to occur again today, there's no way — there's no way — anyone could prove it was arson," he told the paper.
The stark contrast in opinions is emblematic of the split in the field of fire investigations.
Also in 2006, The News-Herald of Ohio reported that Ralph Dolence, a retired investigator then teaching at the Bureau of Alcohol, Tobacco and Firearms, dismissed the "new" arson science as wild nonsense, claiming that investigators like Lentini "muddy the waters" of good investigations. Lentini issued a swift, no-nonsense rebuttal. Still, such accusations seem well past their sell-by dates.
Yet several key players say that finally divesting the shrinking faction of holdouts of their misguided faith in arson indicators like "crazed glass," "alligator blisters" and "concrete spalling" (see sidebar) remains a work in progress.
"I think it certainly is still a problem; I don't think it's as widespread as it used to be," said Bobby Schaal, a 22-year veteran of the ATF speaking only in his capacity as incoming first vice president of the International Association of Arson Investigators. But investigators, lawyers and scientists say that old investigation errors are still being made, and old myths are still being heard in courtrooms.
Common themes emerge when discussing reforms. One hot topic is the routine dependence on negative corpus evidence —simply put, investigators rule out electrical faults and exploding coffee pots, for example, rather than rule in evidence of how a fire did in fact start. So rather than a more accurate description of "cause undetermined," fires are often called arsons based on investigation by exclusion.
To veteran investigator Patrick Kennedy, that practice is unethical and immoral. "I don't know what it is, so it must be arson?" he said. "That is a pretty poor reason to kill somebody."
Prosecutors, defense lawyers and judges could help matters by becoming a safety net of sorts and approaching findings of arson with a healthy dash of skepticism. "I think that the more educated attorneys become and the more honest attorneys become, they can vet out these bad experts," said Schaal, who believes that once they're informed, seeing terms like "spalling" or "crazed glass" in a report should immediately raise a red flag.
To keep flawed evidence out of courts, Hurst also advocates implementing a system of "devil's advocate peer review" in which experts review arson-related evidence pretrial, probably volunteers working via existing regional committees. "If (investigators) start getting their cases rejected," he explains, "or if they're forced back to the drawing board, it will make them much more cautious in making those cases in the first place because they want it to pass peer review."
Trial transcripts and documents could be scanned by computer for distinctive terminology, like "alligator charring." Hurst predicts that with computer scanning, "You're going to find cases like Ed Graf's where old wives' tales will just pop up one after another, just bang, bang, bang." (He recently reviewed the 1988 case of Ed Graf — convicted of burning his two stepsons alive in a shed — for Walter Reaves and declares it "a real stinker, another Willingham.")
The most contentious areas of reform pit the science-qualified experts against the true believers in the superiority of on-the-job experience. "If I have my dream," Patrick Kennedy said, "some day, just like lawyers, you won't be able to enter a courtroom unless you've got a license and a college degree."
And those who embrace science are looking to policymakers and professional societies to raise the industry bar on training, certifying and licensing fire investigators. "I think that the IAAI, the NFPA, the NAFI (National Association of Fire Investigators), anybody should call for the proper credentialing of fire investigators," said the IAAI's Schaal. Lentini suggests all investigators should be certified not by their employers but by a neutral third party like the IAAI or NAFI. And right now, many U.S. fire investigators don't have either of those national organizations' certifications.
Unlike firefighters and public sector investigators, private fire investigators in most states need licenses, and those licenses are covered by private investigator statutes. "They are bogus," Lentini said, "but convey a state credential to someone who has demonstrated a clean criminal history, has filled out a form and sent in a set of fingerprints along with an application and some money. No state license currently requires that a licensed fire investigator actually have any skills in fire investigation."
Practical Experience Plus Diplomas
Certifications and licenses aside, there is a school of thought that only scientists or the scientifically trained should make arson determinations. To understand concepts like thermodynamics and heat transfer, investigators need math, chemistry and physics.
"Fire investigators are largely ex-cops, military, firemen, that sort of thing," Hurst noted. "They don't feel like they need any of this fancy science stuff." He worries that, "no matter how clearly you try to explain these scientific principles to people who are not used to thinking scientifically, you cannot do it."
Brad Hamil, a veteran investigator who long worked with California's San Bernardino Fire Department and is currently second vice president of the California Conference of Arson Investigators, has a different perspective. He greatly values law enforcement training and experience but, "I would say that, yes, there has to be some science in your work," he said. He'd advise newcomers to arm themselves with a bachelor's degree. They won't be any smarter about fire, he said, "but when you're on the stand, 'Well, I have a B.S. degree in science.' Cool. Sounds good to the jury."
Lentini said we've got to live with the fire investigators we have while ensuring the next generation is properly trained. "Certainly, the state legislators should say that people who are going to investigate fires should be qualified by education to do it, and be qualified by experience. They hire firemen," he said. "Eventually, everyone who does this work should be able to document post-secondary education in chemistry and physics, at least at the freshman (101) level."
Required, science-based education may be closer than many realize. If the latest version of the standard professional requirement for fire investigators, NFPA 1033 — compliance with which is required for both the national organizations' certifications — is accepted by local jurisdictions or fire departments, it could go a long way toward keeping unqualified contenders out of the fire investigation business.
Its Trojan horse message, a list of science topics in which an investigator needs training to qualify, could be, said Patrick Kennedy, who sits on NFPA committees, a big surprise to the non-science-based folks who will find that they can no longer qualify under 1033.
"They're going to have to have continuing education in those topics and be able to prove it," he said. He makes no apology for raising the bar. "The problem of people losing their homes or livelihoods in a fire, then being accused of arson with no good reason is," Kennedy said, "the cancer on our industry, and that has to go."
Lentini, who had hoped to see a requirement in 1033 for nonemployer certification within five years, sees progress: "When fire chiefs see that, they're going to say, 'Shit, this means I need a college grad, or at least I need someone who's got some college in science.'"
Until requirements for fire investigators rise across the board, however, results will be mixed. Hamil worked a case with an investigator who made a wrong call on a fire cause. After he admitted to having had only one class in investigations, Hamil reached out, offering to arrange some training options. "I said, 'Call me, I'm the roundtable coordinator for the California conference of fire investigators in my area. We put on training all the time,'" he recalled. "And the phone, to date, has not rung."
Douglas Carpenter is pushing ahead on mapping out a concept for his vision for a national curriculum for fire investigation, ideally a four-year course. That would be expensive. But cities and states can get creative about bringing more science into their fire investigation services.
In December, the Cincinnati Enquirer reported that Cincinnati Police Chief Thomas Streicher Jr., who has a Chief's Scholar Program, sent three officers on scholarship for a one-year master's program at the University of Cincinnati. Surely, appropriate candidates in fire investigation jobs could be put in similar programs. Other affordable options include the IAAI's regional training and online programs at CFITrainer.net, some of which are free.
The fire investigation field is ablaze with ideas despite a dearth of mechanisms with which to make them a reality. And moving forward, many will be looking to the Arson Screening Project for progressive strategies. James Doyle can only aspire to the achievements of the New York Innocence Project. Meanwhile, possible innocents, like Louis Taylor, caught up in flawed determinations of arson are never far from his mind. "Thought has to be given, I think," Doyle said, "to doing something besides saying to prisoners, 'Well, you were born too soon.'"
Are you on Facebook? Become our fan.