So-called "Super Recognizers" are able to recognize a face that they have seen before only briefly. Yet in my surveys I have found that self-described super recognizers cannot give descriptions of faces of even their closest relations even though they would instantly recognize them. I thought this might be due to super recognizers recognizing based on a holistic analysis of the face—a gist—and studies do show this. Others by contrast analyze a face by focusing on the individual components: the nose, the ears, the eyes, whatever else is in a face.
Recognizers can efficiently solve the decision problem of facial recognition but not the computational problem. Given a face they can determine quickly and accurately whether or not they have seen it. Others given a face will either take longer or not be sure whether they have encountered it previously. But the computational problem is describing the face given a name—or even a face. Recognizers—I theorize—fail, while others have more success. To report a face you need to describe each individual component which the recognizers never consciously or even subconsciously process. You cannot draw a gist! Recognizers have a knowledge of the face which is beyond the linguistic.
What is going on in the brain I think—at least metaphorically—is that in recognizers the modules of the brain are operating in parallel while in others they operate in serial. Recognizers see everything at once, the relations between each component. For whatever reason they do not attend to individual components of the face. There is no nose alone, it is always connected to the eyes, the mouth. Other people see a nose, then a mouth, then the eyes, et cetera. There is a sort of dictionary that is indexed by faces in the recognizers but indexed by components in others. One face to one person in recognizers but one nose to many people, one set of eyes to many other people in others. These other people see one face and it could belong to many people but to a recognizer they plug it in and come out with a name. Very few false positives.
All of this is to say that believe me when I say that Roddy McDowall is virtually identical to Burt Brinckerhoff:
A while ago I made a note to myself:
TAKE YOUR PIPE AND SMOKE IT, RODDY MCDOWALL AND BURT BRINCKERHOFF ARE ONE AND THE SAME. THIS IS IMPORTANT…
A Google Image search of Burt Brinckerhoff comes up with no pictures of him like that I posted above, which I took from a Naked City episode he appeared in via YouTube. Anybody looking up the two actors would never come to the conclusion that they resemble each other based on the results of an internet search. And yet I am almost certain that when I wrote that quoted line a Google search revealed their similarity. Somebody scrubbed the results. I was correct, this is indeed important.
Sometimes it is the opposite:
Super-recognizers will not be fooled, but others… lights out!