Eyewitnesses make mistakes, sometimes tragic mistakes. The number of innocent persons convicted, jailed, and maybe executed based on mistaken identifications is unknown, but is considerable. For example, in the more than 250 convictions overturned because of DNA evidence, 73 percent were attributable to misidentifications by eyewitnesses.
Misidentification by eyewitnesses is the largest single cause of wrongful convictions, and is responsible for more wrongful convictions than all other causes combined. As noted by the International Association of Police Chiefs in its 2006 training guidelines, "Erroneous identifications create more injustice and cause more suffering to innocent persons than perhaps any other aspect of police work."
The issue of eyewitness identification is gaining new traction, marked by the recent highly publicized decision of the New Jersey Supreme Court (State v. Henderson), and an important identification case to be argued in the U.S. Supreme Court this term (Perry v. New Hampshire). The New Jersey decision closely examined the more than 200 published scientific studies on human memory and eyewitness identification which through staged events and mock witness experiments demonstrated how human memory is impaired by distortion, contamination, and even false imaginings.
The New Jersey decision explores how systemic variables like suggestive police lineup procedures and suggestive interviewing techniques by police undermine the reliability of a witness's identification. Also noted by the court are the external variables that produce mistakes, such as lighting conditions, distance, stress, presence of a weapon, memory decay, and cross-racial identifications.
In addition, eyewitnesses typically select a person from a lineup who most closely resembles the witness's memory relative to other persons in the lineup, and if the actual perpetrator is not in the lineup, the witness usually will choose the best look-alike. Thus, without external proof that corroborates the witness's identification, one can never know for certain which identifications are reliable, and which are not.
The problem of mistakes is complicated by other factors. Juries typically believe that memory works like a video recording, and that a witness's testimony about the event is simply like replaying the tape. But human memory doesn't work like that. Indeed, there is virtually unanimous scientific consensus that a person's acquisition, retention, and retrieval of information is subject to decay, contamination, and distortion.
Moreover, the most dramatic moment in a criminal trial -- and which has a powerful impact on juries -- is when the witness points to the accused and declares: "That's the man!" Moreover, since the witness is absolutely convinced of her identification, it is exceedingly difficult for defense counsel to cross-examine the witness effectively.
The tools of cross-examination work best when the witness may be lying, and cross-examination is the classic legal weapon to bring out the truth. But the eyewitness's veracity is not being challenged; it is the witness's accuracy that is being challenged, and if the witness has no motive to lie, and is convinced of her identification, such testimony is difficult to undermine.
Courts, prosecutors, and especially police departments have made commendable progress in trying to minimize suggestive influences on eyewitness identifications. Police increasingly use "blind" administrators of lineups who know nothing about the case to ensure that investigators do not convey, even inadvertently, cues to the witness. Instructions are increasingly being given to witnesses by police that the perpetrator may or may not be present in the lineup.
Sequential lineups -- both with live persons and photos -- are increasingly used instead of serial lineups, so that the witness does not know how many persons he or she will view, and there is less opportunity for relative judgments. And lineups are increasingly being recorded to preserve any evidence of improper police conduct.
The courts are also making various efforts to respond to the dangers of mistaken identifications. Courts have formulated special cautionary jury instructions emphasizing the fallibility of an eyewitness's identification, and focusing on the well-documented factors that undermine confidence in the reliability of the identification. Moreover, courts are increasingly allowing experts to testify to the kinds of factors that have been found by scientific research to contribute to mistakes. And the New Jersey Supreme Court adopted a new legal framework for trial courts to ensure that all relevant factors that contribute to flawed identifications are explored at a pre-trial hearing.
These reforms are important and may serve to improve the integrity of eyewitness proof at trials. But again, given the subjective constructions, distortions, and contaminations of human memory, there is the ever-present danger that no matter how protective the system tries to be, witnesses will still make mistakes, and unless there is other external proof to support the witness's identification, one can never know for certain whether the witness has made an accurate identification or false identification. Thus, given the widely acknowledged fact that the eyewitness is the most unreliable of witnesses, there needs to be a new legal approach to the problem.
So, I propose that if a prosecution of a perpetrator rests exclusively on the identification of a single eyewitness, and there is no other external evidence to back up the witness's identification, the prosecution cannot go forward. There is too great a risk that an innocent person will be convicted. The reforms that courts and police are making, although useful, are in many cases simply tinkering with an identification process that is imperiled by serious and unacceptable risks.
The need for corroborating proof to support a prosecution is not a novel or unrealistic idea. The legal system in several instances requires corroborating evidence before a prosecution can proceed. Such corroboration typically is required when the proof by itself is unreliable, and it is only when such proof is reinforced by other supporting evidence that the case is considered sufficiently reliable to proceed to trial. Thus, in some jurisdictions, the testimony of an accomplice to a crime needs to be corroborated.
Also, the unsworn testimony of a young child typically needs to be corroborated. So does a defendant's confession to a crime. It needs to be corroborated by proof that the crime actually occurred. And, a statement from an absent witness that exculpates a criminal defendant also needs corroboration.
Identification of a perpetrator by a single eyewitness without any additional proof has produced more miscarriages of justice than the testimony of any other type of witness, or the use of any other type of evidence.
Courts and legislatures should acknowledge that uncorroborated eyewitness testimony is inherently untrustworthy, and undermines confidence in the integrity of the justice system. Despite the efforts of courts, prosecutors, and police, mistaken identifications produce too many errors, and result too often in innocent defendants being prosecuted, convicted, imprisoned, and even executed.
If corroborative proof is legally required in other instances where the danger of weak or unreliable evidence is great, then it seems only reasonable to require corroboration when the stakes are just as great, and where the documented errors are so well-known and consequential.