The very fact that we are led, in terms of ordinary language, to define as «schematic» the image in figure 3.1 tells us that the image can be supplied as an interpretant and mentally retained as a «model» for mice of different colors, sizes, and (were we capable of discriminating them) individual physiognomic features. And note well that this would happen even if the Encyclopedia contained a photograph instead of the schematic drawing: we would start off from it by operating in a way comparable to the process known as solarization, which is simply a form of deprivation or diminution of the individual features in order to arrive at a rule for the construction of the image of any mouse. And the same thing would happen if, as a result of psychological matters that have to do with the mysteries of the black box, we were to react to the word mouse by calling up the image of that mouse, which we have seen for the first time. The mental representation of that individual would serve us as a cast or model (a schema, in fact), and we would easily be able to transform the experience of an individual mouse into a general rule for the recognition or construction of mice.
However, we can recognize or identify not only natural or artificial objects but also tokens of geometrical figures such as the triangle, and above all of actions and situations (from walking to going out to dinner). If for the type triangle one can think of prototypes or rules for the construction and identification of the figure (not unlike what can happen with regard to the morphological features of a mouse or a glass), and if even notions such as room or restaurant can presuppose a basic visual structure, on the model of Minsky’s frames, then the recognition and identification of actions such as going out to the restaurant, arguing, and scolding, or of situations such as a pitched battle, a meeting, or a sung Mass, require genuine «scripts» (like those proposed in Artificial Intelligence, or representations through Cases and Actants, or more complex narrative sequences, such as Greimas’s schema for anger).25
And that’s not all. I maintain that CTs also include pairs of oppositions: not only is it difficult to interpret husband in the absence of the notion of wife (and we shall be talking about cognitive types by functional genera later) but also in some way a part of our idea of dog derives from the fact that this animal barks or snarls and does not meow or purr (features sufficient for us to decide at night, in the dark, which animal is scratching at the door).
In such cases we undoubtedly possess CTs that do not necessarily or particularly take morphological features into consideration. Likewise we can have CTs that take account of temporal sequences, or logical relations that, while they may be expressed in diagrammatic form (which on the level of expression assumes the form of a visual configuration), still do not regard visual experiences.
If there is a «strong» element in the Peircean theory of the interpretant, it is that the series of interpretations of a sign can also assume «iconic» forms. But «iconic» does not necessarily mean «visual.» Sometimes the CT includes perceptual primitives or even qualia, which are not easy to interpret but which must be accounted for: part of the CT of the skunk—even for those who have never seen one—should include the powerful smell it can give off, and its NC should include the instruction that the skunk is identifiable mainly through its smell (if there were a Kantian schema for skunk, as Kant presupposed one for dog, it would have the form of a flowchart whose higher nodes contain instructions to proceed immediately to an olfactory check).
Are we sure that our CT of the mosquito is fundamentally made up of morphological features and not (eminently) of the irritating effects it can have on our epidermis? We know very little about the form of the mosquito (unless we have observed one under a microscope or seen one in an encyclopedia), but we perceive it first and foremost through our hearing as a flying creature that produces a characteristic whine as it comes closer, and so we can even recognize one in the dark—indeed it is by referring to these features that we would provide someone with instructions for the identification of mosquitoes.
I maintain that the CT (and the NC) of the mouse also includes «tymic» (see Greimas-Courtes 1979: 396) elements. We have already seen how it is fundamental to perceive the mouse (usually) as a repugnant little creature. Apart from its morphological characteristics, the CT of the mouse also includes a frame, a sequence of actions: with the exception of those who have never seen mice outside a cage, the idea of the mouse (and the capacity to recognize a mouse) is based on the fact that it usually appears to us as an indistinct form that moves from one side of a room to the other at high speed, emerging from one hiding place only to slip into another.
This lends particular conviction to the idea proposed by Bruner (1986, 1990) that we use narrative schemata to organize our experience. I think that our CT (and NC) of the tree also includes the (narrative) sequence that it grows from a seed, goes through various phases of development, modifies itself with the changing of the seasons, et cetera. A child soon learns that chairs are not planted but constructed, and that a flower is not constructed but planted. Our cognitive type of the tiger includes not only that it is a big yellow cat with a striped coat but also that if we were to meet one in the jungle, it would behave in a particular fashion toward us (see in this regard also Eco 1990, 4.3.3).
With regard to expressions such as yesterday and tomorrow, can we really say we have only NCs that may be expressed propositionally and not also a sort of diagram with vectorial pointers, the result being that (even if the disposition varies according to the culture) in one case we configure a sort of mental image of «pointing backward» and in the other of «pointing forward» ?
I shall freely adapt a fine mental experiment found in Bickerton (1981). Let us suppose that I have been interacting for a year with a very, very primitive tribe, whose language I understand only very poorly (names of objects and elementary actions, verbs in the infinitive, proper nouns without pronouns, etc.). I accompany Og and Ug on a hunt: they have just wounded a bear, which, bleeding, has taken shelter in its cave. Ug wants to follow the animal into its lair to finish it off.
But I remember that a few months before, Ig had injured a bear before following it boldly into its lair, where the animal demonstrated that it was still strong enough to devour him. I should like to remind Ug of that precedent, but to do so I have to be able to say that I recall a past event, and I do not know how to express either verb tenses or doxastic operators such as I remember that. And so I restrict myself to saying Umberto see bear. Ug and Og obviously believe that I have spotted another bear, and they are frightened. I try to reassure them: Bear not here.
But the pair only draw the conclusion that I make jokes in the worst possible taste at the worst possible time. I persist: Bear kill Ig. But the others reply: No, Ig dead! In short, I should have to desist, and Ug would be lost. So I fall back on nonlinguistic interpretants. On saying Ig and bear, I use a finger to strike my head, or my heart, or my belly (according to where I presume they locate the memory). Then I draw two figures on the ground, and I designate them Ig and bear. Behind Ig’s back I draw images of lunar phases, hoping that they understand the meaning of «many moons ago,» and in the end I once more draw the bear that killed Ig.
If I try this, it is because I presume that my interlocutors have notions of recalling, and above all a few CTs (that may be interpreted not propositionally but diagrammatically) for activities of «protensity» toward temporal points different from the present. That is to say, I start from the principle that, if I am capable of understanding a sentence in which I am told that something happened yesterday or will happen tomorrow, I should have a CT for these temporal entities. In my experiment I would try to interpret visually (vectorially) my CT,
