This piece is the second in a series on defects in our criminal justice system that lead to wrongful convictions. The first piece introduces these seven critical breakdowns in the rules of constitutional criminal procedure.
It’s natural to trust that our memories are usually accurate, since memories are the source materials by which we understand ourselves and the world around us. We revisit them regularly, not just for the pleasure of reminiscing, but for guidance as we navigate our way through life. We treasure some memories—and we’re haunted by others. In the end, sometimes they are all we have. To distrust one’s own memory, especially regarding the important moments of our lives, is a kind of self-abnegation. If we don’t believe our own memory, then what is real?
When jurors hear witnesses testify, under oath, that they watched a defendant on trial commit a horrific crime, that they’re completely certain of what they saw and that they’ll never forget the face of the man holding the knife or squeezing the trigger, jurors are overwhelmingly likely to believe them. The trust we have in memory is even stronger when the witness is also the victim of a violent crime, for whom jurors understandably feel tremendous sympathy. The problem is that we now know this trust is frequently misplaced: Eyewitness misidentification is the leading cause of wrongful convictions. It has sent countless innocent defendants to prison—and left the actual perpetrators free to re-offend.
The Trouble with Memory
Our understanding of human memory has changed dramatically since the onset of the DNA revolution in criminal forensics in the 1980s. “[M]emory does not work like a videotape recorder. You don’t just record the event and play it back later,” explains psychologist Elizabeth Loftus. We don’t recall perceptions stored in our minds; we “reconstruct” them. “Our representation of the past takes on a living, shifting reality,” Loftus says. A memory is like a canvas on which, after the original picture appears, we’re forever painting and erasing.
The first significant study of wrongful convictions published in the United States was Edwin Borchers’ “Convicting the Innocent,” released in 1932. It profiled 65 exoneration cases of various types of crimes from around the country—and 29 of them were the result of witnesses identifying innocent defendants. The Innocence Project reports that as of April 2020, “69% of DNA exonerations—252 out of 367 cases—have involved eyewitness misidentification.” Before he was exonerated by DNA testing in 1993, Kirk Bloodsworth spent years on death row for allegedly raping and murdering a nine-year-old girl, and he had been misidentified by no fewer than five different eyewitnesses.
Usually, witnesses testify in good faith, and so they’re often devastated when they learn that their testimony helped imprison someone innocent. This was certainly the case with Jennifer Thompson, who, while being attacked, made a conscious effort to remember her rapist’s face, only to pick a completely innocent suspect out of a lineup and again identify him as the perpetrator during his trial. Witnesses are often completely honest and devastatingly erroneous at the same time.
External influence, especially in the form of police suggestiveness, can direct how memories are reconstructed. When officers nudge witnesses in the direction of particular suspects—whether unconsciously or, as we’ll see below in the case of Barion Perry, unavoidably—during identification procedures, it can spell disaster. Memory gaps and vague perceptions, as when a terrified victim tries to remember a face when he or she was naturally focused on a weapon, often cause the victim to superimpose a face suggested by police during a lineup onto the attacker. And police routinely implant false memories, often without trying.
To some extent, the Supreme Court has always been aware of this problem. In U.S. v. Wade, decided in 1967, the court recognized the impact eyewitnesses have on juries and ruled that (some) police lineups were “critical stages” of criminal prosecutions that often determined the outcome of trials. “The lineup is most often used,” the court wrote, “to crystallize the witnesses’ identification of the defendant for future reference.” That is, police are already fairly certain prior to the lineup that the suspect is guilty. The purpose of the lineup isn’t so much to help the police find the culprit as it is to lock the suspect’s face into the memory of the witness so he or she will testify confidently at trial. Such confident testimony is often enough to doom a defendant.
Suggestive lineups steer witnesses toward the selection of one participant by making him or her stand out, which happened in the case of Foster v. California. In this case, police arranged a lineup in which Walter B. Foster was taller than the other participants by about a half foot. To expose and deter suggestive police lineups—that is, to avoid police implantation of false memories—the court ruled that the Sixth Amendment guaranteed suspects the right to have an attorney present. However, the court limited its ruling to lineups conducted after the suspect has been indicted or otherwise formally charged by a judge. At this stage in the process, the accused has transitioned from suspect to defendant and the need for a lawyer is especially “critical” because the government, having already zeroed in on a particular individual, is conducting the lineup in preparation for trial.
However, most identification procedures take place pre-indictment, and for that reason, no defense counsel is present to guard against police misconduct. In Stovall v. Denno, also decided in 1967, police, fearing a stabbing victim might die at any time, brought a suspect to the victim’s hospital room for a “show-up.” Show-ups are one-person confrontations between witnesses and suspects, often conducted on-scene and impromptu rather than at a police station. In this case, the Black suspect, Theodore Stovall, was surrounded by five white police officers and handcuffed to one of them when the victim identified him. Show-ups are notoriously suggestive and unreliable because they point a finger directly at the suspect. Witnesses inevitably and rationally believe that the officers seized the suspect for a good reason. A number of courts have ruled that show-ups are “inherently suggestive.” In Stovall, the police could hardly have announced more clearly that they’d found the bad guy if they’d put fangs on him and splashed blood on the front of his shirt.
Theodore Stovall had gone before a judge prior to being identified. However, because the show-up took place prior to the court’s ruling in Wade and the court ruled that Wade should not be retroactively applied, Stovall had no Sixth Amendment right to have a lawyer present, and the court treated the case as if the show-up had occurred pre-indictment.
In Wade and Stovall, the suspect’s protection against suggestive identification procedures derives from the Fourteenth Amendment’s due process clause, which guarantees fundamental fairness throughout the criminal process. Regarding Stovall, the court ruled that the victim’s identification, while obtained during manifestly suggestive procedures, could nonetheless be used against him at trial because the process was not “so unnecessarily suggestive and conducive to irreparable mistaken identification that he was denied due process of law.” The key word in this sentence is “unnecessarily”; the admissibility of certain identifications, no matter how egregiously suggested, the court decided, can be redeemed by exigent circumstances.
Wade and Stovall were decided during the heyday of the Warren court’s criminal procedure revolution, when concern for the rights of the accused was at its zenith. But since 1967, the court has, for the most part, continually enhanced the power of law enforcement officers. Freshly stocked with five appointees from Republican presidents, in 1977 the court extended the power it granted the government in Stovall and a subsequent case, Neil v. Biggers, to also include testimony derived from procedures in which police made no effort at all to minimize suggestiveness. In Manson v. Brathwaite, the court decided that identifications from unnecessarily suggestive processes could be admitted, no matter how brazenly police manipulated witnesses, if, after applying a complicated five-prong reliability test that examined the “totality of the circumstances,” trial court judges believed there was not “a very substantial likelihood of irreparable misidentification.” These reliability screenings would occur during pretrial hearings, outside the presence of juries. Judges would arrive at their decisions after considering, for example, how clearly the witness could see the perpetrator, and how much attention the witness paid to the perpetrator during the commission of the crime—factors that require analyses that are to some degree inexact and subjective. Exoneration data show that in an alarming number of cases, the failures of this filtration process have been devastating.
The qualifying language in the Brathwaite standard typifies the shift from the civil libertarian vision of the 1960s Warren court, which tightened judicial oversight of police lineups in Wade, to the law enforcement-friendly Burger court of the 1970s almost perfectly. A mere likelihood of misidentification isn’t enough to keep the identification from the jury. The likelihood must be “very substantial.” Moreover, a substantial likelihood of mere misidentification isn’t enough. The misidentification must be “irreparable.” Justice Thurgood Marshall, who had battled the government in criminal cases as a lawyer, seemed to intuit that the wrongful conviction floodgates had been opened. Dissenting, he wrote, “In my view, the Court’s totality test will allow seriously unreliable and misleading evidence to be put before juries. Equally important, it will allow dangerous criminals to remain on the streets.”
By the time the court revisited its ruling in Brathwaite in 2012, the continued pace of DNA exonerations had thoroughly exposed the “suggestive lineup-to-wrongful conviction” pipeline. Data on DNA exonerations presented in briefs to the court arguing for reform were compelling and often coupled with harrowing tales of freed inmates whose lives had been broken and scarred. In 2012, instead of overruling or minimizing the harmful effects of Brathwaite, the court ruled in Perry v. New Hampshire that certain suggestive and unreliable identifications are allowed to be used by prosecutors without even subjecting them to the weak and deferential totality test required by the Brathwaite decision. While one might be tempted to give the court the benefit of the doubt for ruling as it did in Brathwaite back in 1977, when crime rates were high and DNA exoneration data didn’t yet exist, the court deserves no such quarter with its ruling in Perry. It was a clear statement that reducing wrongful convictions wasn’t a high priority.
In Perry, a police officer interviewed a witness in the fourth-floor hallway of an apartment building about suspected car burglaries that had just occurred in the adjoining parking lot. Prior to the interview, the witness had seen a different police officer standing next to the suspect, Barion Perry, in the parking lot while looking down from a window. When the interviewing officer asked the witness to describe the perpetrator’s appearance, she responded that he was “a tall black man.” When asked for a more detailed description, the witness gestured to the window and, according to the interviewing officer’s later testimony, said the perpetrator “was the man that was in the back parking lot standing with the police officer.” Though unwittingly administered, this type of show-up, involving a suspect standing next to a police officer at the crime scene at 2:30 a.m., undeniably carries a high risk of implanting false images in the reconstructed memories of witnesses.
While the circumstances of Barion Perry’s identification may have been less suggestive than the hospital room confrontation in Stovall, involving the Black defendant cuffed and standing next to five white officers, subsequent facts suggest the particular identifying witness in Perry was actually far less reliable. Four weeks later, she could not identify Perry as the perpetrator in a photographic lineup. The prosecutor didn’t put her on the stand at trial. However, the officer was allowed to describe her identification at the apartment complex. The jury convicted Perry.
The court’s path should have been clear: When the suspect is a stranger to the victim (as opposed to a familiar face), all identifications derived from gratuitously suggestive identification procedures should be categorically excluded from trial. However, the court ruled in the opposite direction. It not only maintained the ineffective screening system it had put in place with Brathwaite, but it also decided that some of the most suggestive procedures and unreliable identifications needn’t even be screened. The opinion, written by Justice Ruth Bader Ginsburg, focused on deterring and punishing bad policing rather than minimizing wrongful convictions and limited Brathwaite pretrial reliability hearings to only unnecessarily suggestive procedures caused by police misconduct. The witness in Barion Perry’s case had looked out the window at the suspect on her own, not at the prompting of police. Because the police had done nothing wrong, the court reasoned, Perry was denied even the right to challenge the identification’s reliability at a hearing.
In 1952, the court ruled that when law enforcement practices become so irrational, arbitrary and indecent and that they “shock the conscience,” the Fourteenth Amendment’s due process clause requires that they be prohibited. When prosecutors seek convictions based on eyewitness testimony known to be unreliable and known to cause wrongful convictions, as the Supreme Court now allows them to, their conduct rises to this high standard.
In Perry, the court expresses faith that rigorous cross-examination by defense attorneys, testimony of expert witnesses on memory and wise judicial instructions will enable juries to recognize for themselves when eyewitnesses are unreliable. “The jury, not the judge, traditionally determines the reliability of evidence,” Justice Ginsburg informs us. It’s been ten years since the court’s ruling in Perry, and the list of exonerees convicted as a result of mistaken identification continues to grow at a regular clip. As Edwin Borchers wrote in “Convicting the Innocent” back in 1932, “[T]he ways of juries are strange.” Stranger, it seems, than Justice Ginsburg and her colleagues were willing to appreciate.