When Associate Justice Antonin Scalia asked the question at oral argument, hearts filled with hope. Well, mine anyway.
Oh, Nino. Why hast thou forsaken me? The decision is in, 8-1 with Justice Sotomayor the lone dissenter. The majority was written by Ruth Bader Ginsburg, for whom identification evidence, for all its many flaws, does not violate due process when it’s unreliable, as its reliability can be tested through the ordinary means.Why is unreliable eyewitness identification any different from unreliable anything else?
— Associate Justice Antonin Scalia, at oral argument in Perry v. New Hampshire,
Yet again, Justice Scalia says what needs to be said, clearly and decisively. Why indeed? Of the many burdens placed on judges in their role as gatekeeper of evidence at trial, one of the most critical is the preclusion of unreliable testimony. And of all unreliable evidence, none is more dangerous than eyewitness testimony.
As it’s always been done.
As it’s so often, too often, failed.
As for excluding it because it’s unreliable, that’s only to smack the police for impropriety. As long as the police took no part in causing identification evidence to be unreliable, the Supremes have no concern for its admission even if it spells the conviction of the innocent.
Sotomayor’s dissent sums up the problem clearly in its opening:
This Court has long recognized that eyewitness identifications’ unique confluence of features—their unreliability, susceptibility to suggestion, powerful impact on the jury,and resistance to the ordinary tests of the adversarial process—can undermine the fairness of a trial.
Identification is different, having “resisted” the ordinary tests of the adversarial process. No matter how it’s attacked, challenged, questions, it remains intransigent. People believe it when someone says “the defendant was the one,” regardless of whether it’s true. Jurors believe IDs. IDs are frequently wrong. Wrong IDs convicted innocent people. We know this to be true.
Our cases thus establish a clear rule: The admission at trial of out-of-court eyewitness identifications derived from impermissibly suggestive circumstances that pose a very substantial likelihood of misidentification violates due process.
To the extent people pretend a trial is a process to seek truth, we know IDs to be well intended and believed to be true when offered into evidence. We similarly know that they are often wrong. What can be more contrary to due process than allowing evidence so dangerous to the truth to be admitted without some minimal vetting for accuracy?
The Court today announces that that rule does not even “com[e] into play” unless the suggestive circumstances are improperly “police-arranged.” It’s different.
And yet, the concern of convicting the innocent isn’t on the table. The only concern to which the Supremes will admit is punishing cops for impropriety. Innocence? Unreliable evidence? Not our job.
The most comprehensible explanation for this sad abdication of responsibility is that it would be a substantial change from what courts have always done. And yet, it would be no change at all. It’s all a matter of how one looks at it.
Identifications have been the mainstay of criminal prosecutions forever. A very scary thought, knowing what we do about their unreliability, about how susceptible witnesses are to come to believe in their identifications after the fact, no matter how wrong they may be. By the time they hit the witness stand, they’re absolutely certain. They look at the defendant with the most serious and sad face, and proclaim, “I will never forget that face.” Except they never actually saw the face in the first place.
To change this situation is to fundamentally change the nature of trials, as few trials could proceed in the absence of an identification. Of course, the rejection of unreliable identifications wouldn’t make that a bad thing, unless you aren’t terribly concerned about who gets convicted, as long as someone does.
But courts have always been the gatekeepers of evidence. It has always been a judge’s job to refuse unreliable evidence, to reject efforts to sneak in stuff that juries find persuasive but lawyers know is untrustworthy. We keep out hearsay. We keep out bad records. We keep out tons of evidence that would otherwise influence the case, but isn’t worthy of reliance.
Why not identification testimony?
Because the Supreme Court says so. Because that’s they way it’s always been. Because if judges started reviewing identifications for suggestiveness without any claim of police misconduct, maybe the prosecution would have a bit more trouble making its case.
Does it matter that innocent people may be convicted based on unreliable identification testimony? As Justice Ginsburg says, that’s how it’s always been, testing identification through the adversarial process. So what if it hasn’t worked in the past. So what if all those death row inmates were found to be innocent despite the identifications. This is the law, and this is how the law does things. This is how the law has always done things.
Why should identification be different?

