Is this type of classification—and we could talk of categorization if we took the Aristotelian predicables as subcategories—essential for the recognition of something? Not a bit. Certainly not for Aristotle, who failed to define the camel satisfactorily (see Eco 1983, 4.2.1.1) but who nonetheless continued to identify and name it correctly; and not even for cognitive psychology, because no one has ever denied that someone is capable of perceiving and recognizing a platypus without necessarily knowing whether it is a Mammal, Bird, or Amphibian.
In a certain sense, with regard to this matter, the perplexity would be greater for Aristotle than it would be for Kant or the contemporary cognitivists. The cognitivists would get around the problem, if need be, by assuming that there is something of the precategorial in perception. Kant managed to transfer dogs and cats to the ranks of the empirical concepts, and their classification into genera and species into the territory of reflective judgment. But Aristotle tells us that when faced with an individual substance, we understand what its essence is (man or cat), and he would have willingly admitted that it was possible for a slave to recognize a cat even if the slave could not express the definition, yet when he must say what the substance is, he can do it only in terms of definition, by appealing to genus and differentia. It is as if Aristotle were to admit that in some way we have CTs but we can interpret them only in terms of an MC (since the knowledge of classifications appertains to the MC).
Unless he wanted to say exactly what we are saying: that to perceive (by applying categories—his) is to move precisely in what is today called the precategorial, and that attributions of life, animality, and even rationality are precategorial. At least in the sense that Aquinas tried to explain it.19 We do not at all perceive differences such as rationality but infer them from perceivable accidents; the result is that we infer that man is rational through exterior manifestations, for example, the fact that he talks or is a biped. And therefore it is the immediate perception of these accidents that comes to be part of the perceptual experience, and the rest is cultivated elaboration.
What contemporary cognitivism calls categories (which would have been the predicables for Aristotle) are, rather, what the natural sciences call taxa, which are embedded into one another from species to genus (or from orders to classes, or from classes to kingdoms). What cognitivism calls basic categories are certainly CTs, while what it calls superordinate categories (as Tool, with respect to the basic category of hammer) are taxa. Taxa belong to a more complex phase of cultural elaboration and are stored in the MC of some particularly gifted speakers (they depend on a coherent system of propositions, or on a given cultural paradigm).
By the way, I would point out that the distinction was already very clear in John Stuart Mill, when he was examining the various naturalistic classifications that in his day were still the subject of heated debate:
There is … a classification of things, which is inseparable from the fact of giving them general names. Every name which connotes an attribute, divides, by that very fact, all things whatever into two classes, those which have the attribute and those which have not … The Classification which requires to be discussed as a separate act of the mind, is altogether different. In the one, the arrangement of objects in groups, and distribution of them into compartments, is a mere incidental effect consequent upon the use of names given for another purpose, namely that of simply expressing some of their qualities. In the other, the arrangement and distribution are the main object, and the naming is secondary to, and purposely conforms itself to, instead of governing, that more important operation.
Since one cannot combat the inertia of language, I too shall adapt myself to calling these classificatory items categories, but let it be clear that they do not contribute immediately to telling us what a thing is. They show how it becomes hierarchically ordered in a system of basic, superordinate, and subordinate concepts.20
Another observation is that if categories (in the modern sense of the term) are taxa, they have nothing at all to do with those primitives elaborated or identified by «featural semantics»—and that by chance have the same name as many categories or taxa, the ones that are usually printed in small capitals, such as ANIMAL, HUMAN, LIVING, ADULT, et cetera. It is a matter for discussion as to whether these primitives are finite in number, whether they function by conjunction or by intersection, but they are not always organized into hierarchies like taxa, even though in some authors they are organized by relations of hypo- or hyperonymy (for this argument, see Violi 1997, 2.1 and 4.1). In fact, these semantic primitives are often assimilable to those I have called semiosic primitives (which some would define as precategorial).
If noticing that something is a body, it flies in the sky, it is ananimal, and has weight are semiosic primitives, then, if anything, taxa come into being as elaborations of such precategorial experiences—at least in the sense of «precategorial» that I have re-signed myself to respecting.
3.4.3 Semiosic Primitives and Verbalization
Wierzbicka (1996), who backs up her hypotheses with a vast recognition of different languages, persuasively maintains the existence of certain primes common to all cultures. In her view, these are notions such as I, Someone, Something, This, Other, One, Two, Many, Much, Think, Want, Feel, Say, Do, Happen, Good, Bad, Small, Big, When, Before, After, Where, Under, No, Some, Live, Far, Near, If, and Then (my summarized list is incomplete). The interesting aspect of this proposal is that it aims at resolving all other possible definitions in terms of these primitives.
Nevertheless, before going on to utilize some of Wierzbicka’s suggestions, I wish to make it clear that I assume these primes with all due caution. To say that these notions are original does not necessarily mean admitting (i) that they are phylogenetically primitive and therefore innate: they can be primitives only for a single individual, while other individuals start from other, different experiences (for example, seeing will not be a primitive experience for someone born blind); (ii) that they are universal (even though I see no obvious reason for denying this; but we have to make a distinction between the theoretical hypothesis of their universality and ascertaining empirically that precise terms for them exist in all known languages); (iii) that by virtue of their being primitives they are not interpretable.
Point (iii) represents a weakness in Wierzbicka’s argument. This fallacy springs from the fact that it has traditionally been assumed that the semantic primitives mentioned in the previous paragraphs—those presumed features such as HUMAN or ADULT that ought to constitute atoms of meaning that may not be split any further—are noninterpretable. But what Wierzbicka calls primes are not like this—even though the author sometimes tends to treat them as if they were. They are not meaning postulates; they are elements of a primordial experience. To say that a child has a pri mordial experience of milk (and so one presumes that as she grows, she will know exactly what milk is) does not mean at all that the child, upon request, cannot interpret the content of milk (see in 3.7.2 what a child does when asked to interpret the word water). It may be that those experiences expressed by the words see and hear are primordial experiences of this sort, but even a child is capable of interpreting them (with reference to different organs).
By not admitting this Wierzbicka reacts emphatically to the opinion held by Goodman (1951: 57), according to which «it is not because a term is indefinable that it is chosen as primitive; rather, it is because a term has been chosen as primitive for a system that it is indefinable … In general, the terms adopted as primitives of a given system are readily definable in some other system. There is no absolute primitive.» Wilkins has already shown us how it is possible, through a spatial and nonpropositional cognitive schema, to interpret and define both high and low, both toward and under, or inside (see Eco 1993, 2.8.3).21
Having made this reservation clear, Wierzbicka starts off with an acceptable criticism of the so-called definitions of dictionary and of encyclopedia. Take the example of the mouse (1996: 340 ff). If the definition of the term mouse is also to allow us to be able to identify the referent, or in any event to have a mental representation of a mouse (just as Montezuma must have imagined what a horse was like), it is clear that a strictly dictionary-type definition such as «mammal, murid, rodent» (which goes back to the taxa of the naturalistic classifications) is
