The term bird has a greater extension than terms such as hen or sparrow, but I would not say that this means we perceive the CT of bird as a superordinate category with respect to that of the hen. That of «animal that flies in the sky with wings» (which is our ingenuous notion of bird) is a semiosic primitive. For some animals we perceive only that property, and we relate them to the rough CT of the bird. For others, on recognizing certain additional properties, we elaborate a CT with a finer grain.
We recognize a CT of the bird on the basis of the features or the procedures x, y; we recognize a CT of the sparrow on the basis of the features or the procedures x,y, z; and a CT of the swallow on the basis of the property and we realize that there are common features not only between sparrow and swallow but also between a sparrow and other animals we recognize as birds. But at first this must have nothing to do with the logical criterion by
Figure 3.5
which we relate the sparrow to the class of birds, even though it is certainly by starting from similarities that we go on to elaborate taxonomies. We are simply capable of recognizing sparrows, swallows, and birds, and if then someone wishes to devote himself to the study of birds in liberty, he will also have a CT for the curlew and another for the skylark. CTs are generous and disordered; some people have a CT for the cat, some have one both for a tabby and a Siamese, and most of the features of the tabby will certainly be shared by the cat. But even though it seems so evident that it is on this basis that we can go on to state that every tabby is a cat, I insist on reiterating that, on the level of a perceptual process, this is still a suspicion, an intuition of identity of properties, and not yet an inscription in a categorial tree.
If a CT is a procedure for the construction of the conditions of recognizability and identification of an object, see figure 3.5, which shows various 3-D models.
There is a 3-D model for the dog or the horse. Nothing prevents us, owing to more specific needs, from constructing a 3-D model for a Labrador and a pointer, or for a black horse or a Lippizaner, just as nothing prevents Putnam and me from going to work in a nursery one day and learning to distinguish elms from beech trees. But at first, beech, elm, and tree are all CTs that should be put on the same level: and each one of us uses the one or the other according to his own relations with the environment, considering himself more or less satisfied.
The observation sentences I have seen a pointer and I have seen a dog are equally useful and pertinent according to the circumstances, even before it has been decided that the category of pointer is subordinate to that of dog. Perceptually the CT of the dog is cruder than that of the pointer, but it is perfectly adequate in certain circumstances; it does not oblige us to choose between a Great Dane and an Irish wolfhound, and we ask no more than that.
In my view, therefore, the discourse on the CTs still has nothing to do with the discourse on a taxonomic-categorial system. The CTs are only bricks for use in the erection of categorial systems.
Nonetheless there are some possible counterexamples. I admit that the experiment I am about to cite could be used both to equate the categories to the CTs and to deny this equation. Humphreys and Riddoch (1995: 34) tell us of a patient affected by cerebral lesions, who, when shown an insect, drew it with not very much realism but certainly in a way that allows us to recognize something very similar to an insect (fig. 3.6). The fact that he drew it tells us he interpreted it and therefore supplied indications for its identification and future recognition; in short, if he did not have one before, he had now constructed a CT. But when the insect was removed and he was asked to draw it, the subject represented it as a kind of bird (fig. 3.7).
Figures 3.6 and 3.7
When the same patient, capable of recognizing an oyster as such (in the absence of the model), was invited to draw it, he represented it with four legs. The authors note that, for the short-term visual memory, we must postulate deposited mental knowledge, whose degeneration compromises the reconstruction of the remembered object. Could the case be interpreted in terms of a disturbance of categorial competence? In point of fact, in the absence of the insect the patient did not draw a chair and in place of the oyster a pencil.
His memory retained a feature of «animality,» and therefore it could have worked its way up from the insect or the oyster to the superordinate category of the animals, and thence back down toward the birds or some other unspecified beast. But if we consider the perception of animality as a precategorial experience, then by retaining only one vague attribute of what he had seen, the patient might have gone to fish out any CT that contained it, thus slipping from one CT to another, as if surfing in the archipelago of the CTs, instead of working his way up from species to genus.
I am not maintaining that previous knowledge or categorial suspicions do not play a part in the construction of a CT—and the case of Marco Polo mentioned in 2.1 confirms this. I am merely supposing that CTs (i) can be constructed independently of an organized categorial competence and (ii) can also be activated independently and even in conflict with such a competence (as will be seen when we retell the story of the platypus in 4.5).
3.7.4 CTs and Prototypes
3.7.4.1 Stereotypes and Prototypes
Can we identify CTs with those that Putnam (1975: 295) calls stereotypes? If we consider Putnam’s representation of the content of the term water (fig. 3.8),
Figure 3.8
we could say that the CT includes both semantic markers and stereotypical information (while naturally the property of being H2O is part of the MC). In any case the CT has the folk nature of the stereotype, and the random blend of dictionary and encyclopedic elements.
But it is perhaps more interesting to make it clear that stereotypes are not what cognitivist literature has called prototypes.
One of the ways in which the prototype is currently understood is that it is a member of a category, which becomes a model for the recognition of other members that share some properties recognized as salient. When invited to define a bird, Tiny Tim thought of the prototype of the sparrow, for the simple reason that this was the bird he was most familiar with. If taken literally, the experiments carried out regarding the identification of prototypes allow one to think that this is the way we all usually behave.
Others are inclined to consider it as more a bundle of features, and in that sense it would be closer to the stereotype. When we think of a dog (unless we live with one on a day-to-day basis), we do not think of a Dalmatian rather than a Labrador but of a mongrel type. When we think of a bird, we imagine a winged biped of average size (let’s say between a sparrow and a pigeon) and seldom (unless we have come straight from the Arabian Nights) something like the roc. This mongrel form varies according to the culture (I imagine that an inhabitant of the South Sea islands might have a CT of the bird that emphasizes the vividness of the plumage more than is the case with us), but it is precisely in the negotiation of a space for common agreement that the CTs happily mongrelize themselves. Let’s think of an animal such as the dinosaur, which we do not know by direct experience but through real prototypes offered us by the Encyclopedia. Even in this case I maintain that the most common CT is a cross between a dinosaur, a brontosaurus, a Tyrannosaurus rex, and various other extinct giant reptiles. If it were possible to project an average of the mental images that each of us has in this regard, we would find ourselves with an animal out of Walt Disney rather than something we see reconstructed in a natural history museum.40
A third version would have the prototypes as something more abstract, a set of requisites that may be expressed propositionally, necessary if we are to predicate something as belonging to a category; and here the ambiguity of «category» crops up again, since in this last case we are already thinking in terms of classification.
3. 7.4.2 Some Misunderstandings Regarding Prototypes
Prototypes have enjoyed and still enjoy vast popularity in psychological literature, but their history is fairly complex, partly because the person who has worked most on them, Eleanor Rosch, has successively changed her mind about their nature. The scholar who has reconstructed the matter with the greatest precision is perhaps Lakoff (1987), and I will keep to his synthesis.
The story of prototypes springs from a series of questions, from Wittgenstein to
