Somehow—a step never explained—police included Mr. Wright in a photo array shown to the victim that same day when the victim was at a hospital. Here is what transpired:
Detective Gibney presented Ms. Cordle with a photo array of potential suspects. He did not tell her who the suspect was or if the suspect was even in the array. He only asked if she could identify anyone. Detective Gibney observed Ms. Cordle examine the array “thoughtfully.” She then became visibly upset and, without hesitation, she identified [Appellant from the array]. Ms. Cordle testified at trial that she “noticed the eyes . . . [t]he eyes is what set it off.”
Id. at *5–6. [Not pertinent to this article, but the witness then did a voice identification with police playing audio from Wright’s Facebook page.]
On appeal Wright proceeded pro se. He challenged the weight of the evidence, in part by claiming that “the evidence linking him to the crime is inconclusive, questionable, and/or irrational . . . [and] the pre-trial and in-court identifications of him as the assailant is [sic] wholly unbelievable given her assailant was wearing a mask, and, thus, she saw only his eyes during ‘predawn darkness.’”
The trial court opinion rejected that, noting an abundance of corroborating proof but adding that “[e]ven if the identification was made from only [Appellant’s] eyes, there is nothing disqualifying about such an identification if found persuasive by the finder of fact.” The Superior Court Panel “agree[d] with the trial court’s sound reasoning.”
So, is the reasoning sound? Are eyes sufficiently distinctive and distinguishable to properly be found “persuasive” by a jury? Any more or less than “I’ll never forget that cheek” or “it was that elbow—I can tell?” The court never asked that question. Perhaps it should have. The appeal in Wright did not focus on whether an “eye”-dentification has any reliability or is inadmissible lay opinion testimony. No science was referenced, and, of course, the opinion does not include the photo array. But the Superior Court Panel seemingly had no problem with the relevance/admissibility of this form of proof:
Simply put, the jury considered the evidence linking Appellant to the kidnapping and robbery at issue. The jury found Ms. Cordle’s identification testimony credible while rejecting Appellant’s witnesses and defense theories. To the extent Appellant requests that we re-weigh the evidence and assess the credibility of the witnesses, including Ms. Cordle, we decline to do so as it is a task that is beyond our scope of review.
Id. at *15–16. The important sentence that passively approves this form of proof is “[t]he jury found Ms. Cordle’s identification testimony credible.”
That is what juries do—decide what proof if any to believe. But the legal issue is different and what could have been considered is whether this proof is just too weak to have threshold admissibility.
To try and answer that question, I contacted experts in vision science and eyewitness identification to ask what they had to say about “eye”-dentifications. No one had data on how much variation there is among human eyes making them uniquely identifiable, but there is science on how questionable an eye-dentification (or any partially obstructed-face identification) is.
Professor Suzanne Mannes of Widener University, shared research on the risk of error in an identification of a person who commits a crime wearing a mask: Experiments reported in 2020 confirmed that the wearing of a face covering, even one as partial as a surgical mask, can reduce accuracy by about 50%; and 2018 research shows that if the perpetrator is masked but the lineup or photo array has suspects with no masks, then the error rate is greater than in an identification procedure with all the photos or people masked.
Thomas Albright is a vision scientist at the prestigious Salk Institute. He elaborated on this and more, explaining that when the victim or eyewitness sees the perpetrator, they “encode” [bring into their brain] a series of features. When the “recall” process [here the photo array] is in a different condition [encoding a masked face, trying to recall from faces with no mask], the risk of error rises.
Professor Albright concluded with the following observations:
If faces were masked in both the crime viewing and the lineup viewing, an identification might warrant a little more credibility (than mask at initial viewing and no mask at lineup). But it is also well known that human visual object recognition performance decreases with loss of information-bearing cues—as any seeing person would intuit—and I would be skeptical of any identification of a masked face. I think you could argue that this kind of evidence fails the first Manson reliability criterion (such as it is), which is intended to address the question of visual fidelity, since fidelity declines with loss of information.
To paraphrase, “less detail in, less reliable conclusion out.”
Professor Jonathan Vallano teaches law and psychology to both university and law school classes. He noted the lack of sufficient research on such identifications and then raised concerns about whether a subsequent identification procedure would produce reliable results:
In a fair lineup, where all lineup members had the same color/type of eyes as the suspect, the eyewitness’s memory for the eyes—even if distinctive—would not be particularly informative of their decision. And in such a scenario, no other facial features would be available from the eyewitness’s memory to determine which lineup member may be the perpetrator, as well as distinguishing between other lineup members who are not the perpetrator.
All three professors also spoke about holistic face processing. As explained by Professor Mannes, “Unlike many other objects, almost all faces contain the same features—eyes, nose, mouth, and so on. Thus, in order for us to be able to identify an individual, the features and their relative position, size, and color must be processed as a package, or as a ‘whole’ rather than as a set of unrelated features.”
Professor Albright added that “viewing of the full face in the lineup engages holistic face processing mechanisms in the human brain, which are defeated because the encoding was only of visual features (e.g., eyes, forehead, hairline).” Professor Vallano concluded that “such a lineup would therefore be an overly difficult test of the eyewitness’s memory, which would therefore seem to generally produce a less reliable result overall.”
Because of the compelling evidence of guilt separate from the “eye” and voice identifications, whether a mistake occurred in Walker’s case by admitting this testimony may be a moot point, one of harmless error. But judicial opinions should proceed carefully and not rubberstamp the admissibility of proof fraught with risk and unreliability. There is too much at stake to permit eye-dentifications, and trial judges should scrutinize whether a claim of “that’s the eye” has any basis in science or human experience.
