Memory is a compulsive liar.
On the morning of Dec. 19, 2004, I was driving on Route 1A in Revere when I passed the scene of a fatal car accident. Within minutes, queasy, heart racing, I was on the phone with one of my colleagues at the Globe, describing everything I remembered: A gray van, upside down, its front crushed.
I got it wrong. The next day’s story revealed that the van was maroon, not gray. It was on its side. And, perhaps distracted by concern for the van’s occupants, I had completely missed the utility pole laying across it.
I’ve been thinking about those faulty recollections, and memory in general, since I wrote last week about Angel Echavarria, who finally walked free last Monday after spending 21 years in prison for murder. His conviction was based solely on the memories of two eyewitnesses, who identified him as one of the men who killed a Lynn drug dealer on a snowy January night in 1994.
If we had known as much about the limits of memory then as we do now, he might not have spent almost half his life behind bars.
Since genetic testing began in the 1990s, an astonishing 73 percent of convictions overturned through DNA evidence were based on eyewitness testimony, according to the Innocence Project at the Cardozo School of Law at Yeshiva University. Fully 38 percent of those cases rested on identifications made by two or more eyewitnesses.
How can multiple people get it wrong? As my own experience shows, they’re not necessarily lying. “Memory is not like a videotape, where what goes in and what comes out is seamless and coherent,” said Rosanna Cavallaro, a professor of law at Suffolk University. “It is much more malleable, and susceptible to influences of all kinds.”
The process of recollection alters our memories. Numerous studies show that squishiness: Subjects who view footage of a fender-bender will later recall it as a high-speed crash if an interviewer suggests it was more serious.
Fear compromises memory, as does bad lighting, and the presence of a weapon: If you’re focused on the barrel of a gun — as were the witnesses in the Echavarria trial — you’re understandably less likely to remember details about the person pointing it at you.
“A witness’s memory is part of the crime scene, like physical trace evidence,” said Boston lawyer James Doyle, an expert on eyewitness testimony. “It is subject to contamination . . . like blood or saliva found at the scene.”
The memories of the witnesses at Echavarria’s trial were treated not like trace evidence, but ironclad proof. Eight days before he identified Echavarria as his brother’s killer, Isidoro Rodriguez had named somebody else. The fact of those two differing identifications would have made his testimony less credible — if the jury had ever heard about it.
A year after the crime, police found a second witness to identify Echavarria. We now know that the way they secured the second ID — using a detective involved in the investigation — made it less reliable. Even a well-meaning officer who knows who the suspect is can give unwitting clues when showing a photo array, overwriting a witness’s memory.
There is no sure way to tell which memories are reliable, and which are not. Which is why we need to focus on how police and prosecutors get those initial identifications. Massachusetts has been a leader here, with DAs and police following rules that reduce the risk of contaminating eyewitnesses accounts.
And in January, the Supreme Judicial Court ordered that jurors be instructed on how memories can be compromised, and how an eyewitness’s sense of certainty may not mean their memories are more accurate.
We are finally starting to get the hang of this, and not a minute too soon. It all comes 21 years too late for Angel Echavarria. How many others languish in prison because of the system’s blind faith in a faculty that can transform, in the blink of an eye, a maroon van into a gray one, an innocent into an inmate?Yvonne Abraham is a Globe columnist. She can be reached at firstname.lastname@example.org. Follow her on Twitter @GlobeAbraham.