If we do not postulate physiognomic types, the fact that we are able to recognize the same person over the course of time remains inexplicable. Year after year people change, the face gets fatter or thinner, wrinkles appear, the hair grows white, the shoulders sag, the walk loses its elasticity. It is prodigious that in normal circumstances we can recognize someone after having lost touch for a great many years. If we do not recognize him right away, all we need is a tone of voice, a glance, to push us toward recognition and the ritual «You haven’t changed a bit!»
This means that we had constructed a physiognomic type of the subject with only a few salient features of the original, which sometimes have more to do with a way of moving the eyes than the shape of the nose or the quantity and length of hair. We memorize a sort of gestalt of the face (or of the posture, or sometimes the gait) that can even resist changes in each individual property.
The extent to which the physiognomic type is schematic is well known to lovers, prone to having two apparently contradictory experiences. On the one hand, they are always under the impression they have spotted their beloved in the distance, only to realize later that they were wrong: this amounts to saying that desire led them to apply the physiognomic type with generosity, trying to make it applicable to many concrete tokens. On the other hand, when the loved one is absent, lovers try desperately to reconstruct his or her features in the memory and are constantly left disappointed by the fact that they do not have the same intense feeling that occurred when they saw their beloved directly. In this case they find out how the physiognomic type serves for the recognition of tokens but not as a surrogate for direct perception of the token (with the exception of subjects with an eidetic memory, like many artists who can draw a portrait relying on memory alone). In other words, they become aware of the remarkable difference between «recognition» and «recall» (see Evans 1982, 8).
But physiognomic types for individuals possess a characteristic that distinguishes them from generic CTs, which, however private they may be, can usually be made public in the form of an interpreted NC. It certainly happens that someone can easily recognize mice but cannot or has never had occasion to express the morphological features by which he recognizes them, and therefore we have no guarantee that with regard to mice this person has a type like that of other people (for idiosyncratic reasons he may recognize them only by their rapid movement, having no notion of their shape).
When we talk about mice to this person, at best he would describe them as «disagreeable rodents that thrive in the home,» and since this notion is a part of the common NC, we would come to the mistaken conclusion that this person’s CT has the same format as our own and shares with our CT a knowledge of all the morphological features that are part of the area of common knowledge. But the circumstances of communal life make a case of this kind highly improbable, and while this could happen nowadays with mice (very seldom seen by the vast majority of people), it could happen only rarely with a cow and extremely rarely with a chair.
The same thing does not happen with physiognomic types of individuals. Note that the phenomenon occurs not only with humans but also, a fortiori, with individual animals, vegetables, and artifacts. Anyone will agree about what a dog, a bicycle, or a pipe looks like, but it is extremely hard to explain to someone what the dog Tom, my bicycle, or my pipe looks like. In the case of animals and objects, the generic features usually prevail, and we sometimes have trouble recognizing our car from among a large number of cars of the same make in the parking lot (unless our car has distinguishing marks). But the import of the problem is different with regard to human individuals.
I would recognize Johnny among a million individuals, and it would be the same for Mark, and yet the reasons why I recognize Johnny may be enormously different from those that lead Mark to recognize him. Mark and I could spend a lifetime referring to Johnny, and both of us would recognize him when we met him, without ever having had occasion to make public the features through which we identify him. We would notice the difference between our CTs only if one day we were both asked to collaborate on his police sketch: only then would I discover that Mark not only has never paid attention to the shape of Johnny’s nose but also has ignored Johnny’s abundance of hair or incipient baldness, and perhaps he considers him slim while I see him as robust.
If someone were then to ask us who Johnny is, on interpreting the content of the name, we would notice not only that our interpretations fail to coincide but also that the boundaries between NC and MC would be very vague. Perhaps both of us might say that he is a human being, of the male sex, a professor of such and such at the university of such and such, but for me Johnny would be Louis’s brother and the author of a renowned book on the Nahuatl language (which was Montezuma’s native tongue), while Mark might show that he did not know these details. And yet just one of these details could enable a third party to associate the name Johnny with a very large number of other properties, and even push this third party to disinter, from his own memory, data useful for Johnny’s identification. For his part, Mark might be the only one to know that Johnny has the property of being Jack the Ripper, and nobody would dream of saying that this is an insignificant property—even though it strikes me as being a part of the MC and not the NC.
Let us say then that with individuals three phenomena occur: (i) the frequently idiosyncratic nature of the CTs that make it possible to recognize individuals, (ii) the difficulty of interpreting these CTs publicly and therefore of providing instructions for identification, and (iii) the elasticity of the properties that may be expressed in terms of an NC. I think that this is one of the reasons many theoreticians maintain that the proper names of individuals have no content but designate their bearer directly. Clearly this is a foregone conclusion, because much of our life is spent defining (for others) the various individuals that we name by correlating their names to an occasionally vast series of properties, expressed through verbal descriptions and visual representations; but it is certain that such descriptions express features that in certain situations and for some people are salient, despite the fact that the features are not always salient for everybody, while there can be a notable gap between one interpretation and another.45
In 1970,1 grew a beard. Twenty years later, I shaved it off for a few months, and I noticed that some friends did not recognize me at first sight when they met me, while others immediately established a normal interaction, as if they were unaware of the change.
I understood later that the subjects in the first category had known me only in the last twenty years, therefore when I already had a beard, while those in the second category had known me before I grew the beard. We construct physiognomic types of the persons we meet (almost always based on first impressions or, in exceptional cases, on the moment in which the impression was most vivid) and rely on that for the rest of our life—in a certain sense we adapt the features of the «new» token we meet to the initial type, rather than correct the type at every new encounter.46
This leads me to think that, just as caricatures emphasize features that are really to be found in the face portrayed, and just as the study of stupidity often serves to gain a better understanding of intelligence, much pathological behavior does little more than emphasize «normal» tendencies, which are usually controlled by and reabsorbed into more complex models of behavior. I am thinking of the studies on prosopagnosia and in particular of the fine analysis made by Sacks (1985) on the man who mistook his wife for a hat. Since not even Sacks knows what really happened inside Mr. P.’s black box, we can content ourselves with considering Mr. P.’s verbal interpretations.
P. does not recognize faces, but he is not just suffering from prosopagnosia, he also has generalized agnosia and does not recognize landscapes, objects, or figures: he focalizes his attention on particular features without managing to compose them in a global image. He gives a minute description of a rose but does not identify it
