Reprinted with permission from the American Bar Association, The Journal of the Section of Litigation.
Some bad ideas just refuse to go away. A couple of decades ago, defense lawyers in criminal cases, looking for some kind of edge against eyewitness testimony, seemed to strike gold when they came upon psychological research showing that eyewitnesses are not what they are cracked up to be. Eyewitnesses, said this newfound science, routinely make mistakes. Far from providing reliable evidence—which these researchers contended everyone supposedly believed was the case—eyewitnesses often get things wrong, skewing the pictures before their eyes with the biases of their brains, not to mention their fears, or anger, or likes and dislikes. Our presumed faith in eyewitnesses, this research claimed, was somewhere between doubtful and delusional.
This seemed just what the doctor ordered. But defense counsel still faced the problem of how to get this new science into a criminal trial. Enter the “witness identification expert.” Although part of an emerging field, these psychological researchers were thought likely to pass muster under Daubert and its progeny as experts who could be helpful to the jury. The idea was not to put these scientists on the witness stand to challenge directly what any particular eyewitness in the case had said, or even comment on any other actual evidence at trial. They would merely alert the jury generally to the perils of relying on eyewitness testimony at all. The hope was, of course, to shake the jury’s confidence, such as it may be, in what any eyewitness might have said and thus create reasonable doubt about a defendant fingered by witnesses as the perpetrator.
Courts didn’t like it. It is not difficult to see why. To begin with, as any devotee of the movie My Cousin Vinny would be quick to tell you, the conclusion that eyewitnesses can be mistaken is hardly startling. Even in a case less compelling than a lady with Coke-bottle-bottom eyeglasses, we all, as everyday citizens, have had ample experience with the idea that we “can’t believe our eyes” or that we were tricked by what we thought we saw. So persuasive is this little bit of simple common sense, it has become a matter of legend and lore. Lawrence Durrell’s Alexandria Quartet, for example, and Akira Kurosawa’s classic movie Rashômon both depend on exactly the principle that witnesses are often mistaken about what they see. Likewise, the development of DNA testing has shown that witnesses frequently get it wrong. Misidentification of a perpetrator is merely a subset of a larger problem. Two eyewitnesses to the same event often give different reports. Indeed, isn’t every trial a contest in part about the different perceptions of the witnesses?
More fundamentally, an expert on witness identification does not really provide evidence at all. He is, instead, challenging the jury’s competence to evaluate properly the evidence presented. Nothing else in a trial is of this character, unless it is the judge’s instruction that the jury is to decide the facts while it is her province to decide and instruct on the law. This exception, by itself, should underscore how highly problematic an expert on witness identification testimony is. It upsets the very assumptions underlying courtroom dynamics. For reasons such as these, the proffered testimony was rejected by trial judges in most cases, and challenges to the trial courts’ rulings, subject to an abuse of discretion standard, largely failed. Typical was People v. Enis, 139 Ill. 2d 264 (1990), in which the Illinois Supreme Court ruled that a trial court’s decision to bar expert testimony regarding witness identification should be upheld.
People v. Lerma
Perhaps because of the pressures on defense lawyers to come up with something, anything, to defend clients fingered by eyewitnesses, the idea has refused to die. And now the Illinois Supreme Court has reversed itself, breathing new life into the idea and making it nearly certain that such witnesses will become a routine feature of criminal trials hereafter.
The eyewitness testimony in its chosen case, People v. Lerma, 2016 IL 118496 (Ill. 2016), could hardly have been more compelling. There were two eyewitnesses to the murder, and little else for the jury to go on. The first witness was the victim himself, whose dying breath was admitted into evidence as an excited utterance. Before he expired, he blurted out that his assailant was “Lucky,” which other testimony made clear was a reference to the defendant. The second witness was the victim’s girlfriend. She was with him on the front stoop of his home when the shooting began and was shielded from a hail of bullets by the victim himself. She not only heard the victim’s dying words identifying “Lucky” but herself also got a glimpse of the triggerman. She identified him as “Lucky” too. Her account seemed to contradict some of her grand jury testimony in minor particulars, and there were some contradictions regarding how frequently she had seen “Lucky” in the past and whether she merely knew him or had met him or the like. But she was confident about who the shooter was.
Enter the expert. “Lucky” Lerma’s lawyers, presumably having little other good fortune to go on, sought to put a psychologist on the witness stand to tell the jury how unreliable eyewitness testimony can be. Not this eyewitness testimony, mind you, but just eyewitness testimony generally. This witness apparently would have nothing to say about any of the actual evidence in the case itself. The prosecution objected, focusing less on the peculiar character of the evidence than on a showing that the expert in question had stated in his scholarly work that his findings were of questionable value where the witnesses knew the person identified. The trial court was having none of it anyway, citing Enis and stating its own view that witness identification experts add nothing to what is already known about the reliability of eyewitness testimony.
Later in the proceedings, a second such expert was proffered, the original expert having passed away. This one made even clearer than his predecessor that he would not be testifying about the specific facts of the case. Unlike the first expert’s writings, the second’s scholarly work expressed the view that misidentifications often occur even when (especially when?) the witnesses know the person identified. The notion is that witnesses might easily associate someone they don’t know with someone they do. The judge was unimpressed by these new wrinkles and again dismissed the idea with a wave of the hand.
Too quick a wave, ruled the appellate court, which found fault with the trial judge’s over-easy assumption that the second expert had nothing to add. According to the appeals court, the trial judge’s failure to look more deeply into the matter once the second expert presented a more compelling possibility than the first was sufficient to qualify as an abuse of discretion. Affirming, the Illinois Supreme Court went further, giving its reasons the trial court should have allowed the expert to testify.
These reasons were thin, to say the least. It found that “the current state of jurisprudence” had changed dramatically since Enis. “[E]xpert testimony concerning the reliability of eyewitness testimony,” it stressed, “has moved from novel and uncertain to settled and widely accepted,” citing a Supreme Court of Pennsylvania case from 2014. The reason for greater judicial acceptance, it explained, was that “although findings of the sort described by [the proffered expert] reports are now ‘widely accepted by scientists,’ the findings ‘are largely unfamiliar to the average person, and, in fact, many of the findings are counterintuitive.’” The court asserted that with further research we now know not only that eyewitness identifications are not always reliable but also why that is the case. Because the “only” evidence of the defendant’s guilt in the case was eyewitness testimony, the expert report should have been admitted.
The sense of this is a little difficult to see. The outcome in Enis and related cases did not turn so much on the reliability of the science as on the still-pertinent question of whether the testimony added anything to the jury’s deliberations at all. Perhaps we do know better now why eyewitnesses may get things wrong. But how does the why add to the whether? Eyewitnesses make mistakes, we all know that. Do we really need to know, or care, why this happens?
To put it somewhat differently, so long as a jury is aware from its own experience and common sense that eyewitnesses are not always reliable, the psychological mechanism giving rise to this defect would seem to make little difference—no more important than why we feel hunger, or want to eat, compared with the act of eating itself. And so the greater confidence scientists have about the “why” changes nothing or adds a mere accent to a simple commonplace.
In short, the court seemed to go off on the wrong track. The real question is not whether the research is sound, but whether it is helpful to the jury, other than in the most oblique and questionable of ways, fortifying a juror who arbitrarily doubted an eyewitness or perhaps even causing an otherwise confident juror to doubt herself. Is the abstract notion that eyewitnesses often get things wrong, without more and without any relationship to the actual eyewitness evidence in the case, what we would call “reasonable” doubt? Isn’t it merely an attempt to have the jurors substitute theoretical doubt for their own commonsense evaluation of the actual evidence?
How to Address the Issues
To appreciate the full character of this evidence, first think about the opportunity to address these issues in a more traditional way, in line with standard practice. An astute cross-examiner, for example, could, by good questioning of a witness, be able to cover everything ever wanted from this expert. Even in Lerma, there were ways to poke holes in the girlfriend’s account. Or a lawyer can argue the point to the jury, appealing to its own experience with misidentification. In short, our system of justice provides a vehicle for lawyers on both sides to challenge or support a witness’s credibility without resort to abstractions that may or may not apply to the individual case. As in Lerma, this kind of expert can be expected to most have an impact in those instances where the lawyer can think of nothing else, often because there is nothing else. It will tend to sow doubt where there really is none, while perhaps exempting a defense lawyer from doing his job of finding something more concrete and viable to go on.
It is not a sufficient response to say, as some defense lawyers might, that a defendant is entitled to use whatever means are available to raise a reasonable doubt. Surely, this is an overstatement. The question is whether the evidence permits a reasonable doubt, not whether something apart from the evidence itself supports some doubt. We have a system designed to protect the wrongly charged at some risk of the guilty going free. But the balance struck by the system likewise permits a jury to make that decision based on its own judgment about the evidence, without simultaneously being told that its own judgement is, or may be, defective.
There is, moreover, a larger problem. Note that this testimony singles out eyewitness testimony, claiming it is unreliable. But, even were these experts correct, is this evidence less reliable than other forms of evidence? One can think of other forms of proof, routinely relied on, that are far more unreliable. Take for example circumstantial evidence. Or even documentary evidence, which, with email, is now all the rage.
Follow this out and you will see there is something more important at work here than might first appear. The U.S. Constitution states that a person may be convicted of treason only upon the testimony of two eyewitnesses or a confession. What’s the reason for this rule in a case of the most heinous of political crimes? The requirement for two eyewitnesses is a tacit admission that a sole eyewitness may be mistaken, acknowledgment in and of itself that the witness-identification expert’s spiel is hardly revolutionary. It then offers a confession as an alternative. This seems to give us a rank order of reliability. Confession is the most reliable. Eyewitness testimony, especially if supported by two witnesses, seems second best. Documentary evidence is not allowed at all (unless perhaps it includes a confession, though that is not clear).
Isn’t this right? Although eyewitness testimony may be less than completely reliable, it is more reliable than most other types of evidence. Why would we allow expert testimony with respect to this one form of evidence without allowing it with respect to the others? After all, might not an expert come in to say that, psychologically speaking, there are all kinds of reasons that a confession may prove to be unreliable? Should we next permit experts to testify about which types of evidence are more reliable or less reliable?
Should we permit experts to explain logic to the jury? Are we to have experts to explain the problem of a “false correlative” in the former case, wax rhetorical about logical fallacies like post hoc ergo propter hoc, that if one thing follows another, it does not mean the latter caused the former? (No, trailer parks do not cause tornadoes, just because a trailer park always seems to be around when there is a tornado.) As for documents, couldn’t an expert note the ways in which docu-ments—which, after all, are the products of eyewitnesses—can be just as unreliable as the eyewitnesses are? Or more so?
Any system of justice is going to depend on some kind of proof, recognizing that no form of proof is absolutely reliable. Ours depends on confessions, eyewitnesses, documents, and circumstantial evidence—probably in that order—all of which, within carefully devised rules, is put to a jury to use its common sense on to decide questions of liability and guilt. That is, our very justice system, while presumably recognizing the limits of eyewitness testimony (as the Constitution in fact confirms), has decided to allow jurors to rely on it. What an expert in this instance is doing is actually calling into question the very basis of our judicial system. No other testimony is like this.
And where might it end? What’s to prevent an expert from commenting on the division between judge and jury? Or questioning the adversary system itself, perhaps a scientist who sows doubt about whether the clash of opposing viewpoints really should be viewed as most likely to have the truth come out?
You get the point. The goal of the eyewitness expert is to attack the system’s very assumptions, not to add to the evidence. It only seemed less outrageous in Lerma because we were protecting a criminal defendant, and so much of that case depended on eyewitnesses. But the principle is capable of extension well beyond the case presented. And it’s best put aside as a bad idea whose time, if ever, has passed.