This insufficiency seems to extend even to the definition proposed by the Encyclopedia Britannica, which starts off with a zoological classification, specifies the areas in which the mouse flourishes, and expatiates on its reproductive processes, its social life, its relations with man and the domestic environment, and so on. Those who have never seen a mouse would never be able to identify one on the basis of this extremely vast and organized collection of data.

In opposition to these two definitions Wierzbicka offers her own folk definition, which contains primitive terms only. The definition takes up two pages and is made up of items of this type:

People call them Mice—People think that they are all of the same kind—because they come from other creatures of the same kind—People think they live in or near places where people live—Because they want to eat things that people keep for people to eat—People don’t want them to live there…

A person could hold one easily in one hand—(most people wouldn’t want to hold them). They are grayish or brownish—One cannot notice them easily—(some creatures of this type are white)…
They have short legs—because of this when they move you can’t see their legs moving and it seems as if their whole body touches the ground…

Their head looks as if it were not a separate part of the body—The whole body looks like one small thing with a long thin hairless tail—The front part of the head is pointed—It has a few stiff hairs sticking out sideways—There are two round ears sticking up one on each side of the head—They have small sharp teeth that they bite things with.
This folk definition recalls the Kantian idea that the schema for dog must contain the instructions for imagining the form of the dog. If we were to play one of those parlor games in which someone verbally describes a drawing and someone else must manage to reproduce it (a game measuring at once the verbal capacities of the former and the capacity for visualization of the latter), the game might succeed, in that the second person could probably respond to

Figure 3.1
the description-stimulus proposed by Wierzbicka by drawing an image like that shown in figure 3.1.
But is the image only the interpretative output of the verbal definition, or is it a primary and constitutive element of that definition? In other words, is this morphological schema also a part of our NC of the mouse? A good encyclopedia should insert, into the long and satisfying scientific definition of the mouse, also a drawing or a photo of a mouse. Wierzbicka does not trouble to tell us whether the encyclopedia she consulted contains an illustration, nor whether it is a bad thing that it does not. This oversight is not fortuitous: the explanation comes on, where it is maintained that language cannot reflect the neural representation of color, because the representation is private, while language «reflects conceptualization.» It is for this reason that, while seeking to gain a firm hold on the notion of semiosic primitives that should precede the very processes of categorization, she ends up by recognizing the primitives only insofar as they are expressible in (general) verbal terms, with the result that the semiosic primitive of «something» is deliberately printed as SOMETHING— in other words, as if it were a semantic primitive closely bound to the use of verbal language.22

3.4.4 Qualia and Interpretation

If there were noninterpretable primes, we should have to return to the problem of qualia (which I thought I had laid aside in the previous chapter) and return to Peirce. Let us pose the problem in its toughest and most provocative form: Do we have CTs for qualia? If the answer is no, then qualia are «bricks» for the construction of CTs, but in that case we can say neither why we predicate them (this thing is red or boiling) nor why we usually agree about such predications—albeit at the cost of some negotiation. Peirce had said as much: the first feeling that I have of something white is pure possibility, but when I proceed to the comparison of two qualities of white, I can begin a series of inferences and therefore of interpretations; the perceptual judgment desingularizes the quality (CP 7.633). This shift to Thirdness is already a shift to the universal.

It has been a matter of endless discussion as to whether my feeling of red is the same as my interlocutor’s feeling, but, except in cases of color blindness, when I tell someone to fetch me the red pen, there is usually a case of felicitous reference, and I do not receive the black pen. Since felicitous reference has been assumed as a proof that the CTs exist (in the black box), then qualia too have CTs.

Yet again I limit myself to saying that they must be there, and I do not allow myself to say how they are constituted. But a good proof that there is a CT is that it can be interpreted. Can we interpret qualia? We can in the sense that I can not only define red in terms of the corresponding wavelength but also say that it is the color of cherries, of the jackets of the Canadian Mounties, and of many national flags. In addition, through various comparisons I can interpret various qualities of red. Finally, experiments on categorial perception (see Petitot 1983) tell us that there are «catastrophe points» on this side of which the subjects perceive red and on the other side of which they perceive another color—and even though the catastrophe point varies depending on the exposure to the stimulus, it varies in a constant fashion for all the subjects.

Sensations of sweet or bitter are private events, and yet wine experts use persuasive metaphors to discern the flavor and consistency of wines, and if they were unable to recognize qualia on the basis of a CT, they would be unable to tell a Pinot from a Tokay, nor would they be able to identify the vintage.23

One of the usual proofs against the interpretability of colors is that they cannot be interpreted for the sightless. All we need to do is agree on what is meant by interpretation. In Peircean terms, an interpretant is that which lets me know something more about the object expressed by the name, but not necessarily that which lets me know everything that other interpretants tell me. It is obvious that someone who is blind from birth cannot have any perception of red, a semiosic primitive that may be acquired only through perceptual experience. Nonetheless let us suppose (and the experiment is not all that far from scientific fact, see Dennett 1991, 11.4) that the blind person has been equipped with a video camera inserted in special eyeglasses, capable of identifying colors and communicating them in the form of impulses to some part of the body: faced with a traffic light, the blind person, trained to recognize different impulses, would know whether it was signaling red or green.

We would have equipped him with a prosthesis capable of supplying him with a datum that would allow him to make up for the missing sensation. I am not deciding whether or not he would «see» something similar to red in his brain, but that his brain would register an interpretation of red. In order to characterize an interpretation as such, it is unnecessary for it to appear as perfect; on the contrary, every interpretation is always partial. To say to the blind person that red is the color of incandescent substances is a vague interpretation, but it is not less satisfactory than telling someone that a heart attack is that thing you perhaps have when you feel sharp pains in your chest and left arm. On becoming aware of a pain in the chest, we have as many reasons to say Maybe this is a heart attack^ as a congenitally blind person has to say Maybe this substance is red on becoming aware of an intense feeling of heat. The congenitally blind person simply registers red as a «hidden quality,» just as we register something that manifests itself in the form of a symptom as a hidden quality.24

3.4.5 The CTs and the Image as «Schema»

If we found something interesting in the Kantian notion of schema, it was not when the schema appeared to us as something extremely abstract such as «number,» «degree,» or «permanence of the manifold,» but precisely when (and it was at this point that the first Critique was unable to provide a satisfactory answer) it had to permit the formation of an empirical concept such as that of dog (and mouse). We saw that it was necessary in some way to introduce the instructions for producing a figure into the perceptual process. The image of the mouse in figure 3.1 must not be seen as the image of one particular mouse (not even if it were a photograph, which could only be of one particular mouse). And in fact when we see images of this kind in an encyclopedia, we do not think they must supply us with visual instructions for the identifi cation of an animal «exactly» the same as the one shown. In point of fact we assume them as images of the mouse in general.

How can we start off from a «picture» (ever fated to be the representation of an individual, even were it the image of a triangle, which can be only the image of a certain triangle and none other) and use it as a general schema with which to identify or