8

REGULARITY AND CONTEXT

8.1. Regularity and variation in language

The set of semantic competences that I have outlined and discussed thus far can be defined as a controlled idealization, or at least a controllable one, in the sense that there are standards on which to base verification criteria and procedures for evaluating the extension and value of such competences. Dictionaries are specific registers that bear witness to the possibility of intersubjective control; the updat­ing of dictionaries, which verify collective agreement on linguistic change, dem­onstrates the possibility of a continual process of diachronic adjustment between inventories and usage.

The meaning attributed to the words of a language by individual speakers is thus regulated by a general, intersubjectively founded convention which, within a given linguistic community, is assumed to have a regularity that underlies use and the inevitable variations in individual competence. Competence, convention, and linguistic community are mutually related terms; the meaning for which compe­tence is presumed is that which is assumed to be shared within a given commu­nity. This presupposes the existence of a regularity of linguistic meanings: we have competence in the standard, regular meaning of a term, which is conventionally delimited by the language.

For lexical semantics, regularity, or typicality, may be the most important idea to emerge from the study of categorization processes. In chapter 5 I suggested a distinction between protypicality and typicality, in order to indicate two cog­nitive modalities for organizing sense that are connected but different, a difference that has not always been adequately underlined. Prototypicality relates to the ex­istence of a more significant exemplar of a category, such as the sparrow or swal­low in the category of birds. As I have already pointed out, this meaning is only oresent at the basic level of the categorial hierarchy; when we shift to the subor

182

dinate or superordinate levels, the sense of "prototype" changes and is no longer "a type of" but the standard case, the average exemplar. If the prototype of bird is a sparrow, the prototype of sparrow is the typical sparrow, the most regular case. This second meaning of prototypicality, which I proposed calling typicality, is more important for lexical semantics, in that the idea of "average value" as typical value can provide a basis for the representation of lexical meaning. The first mean­ing, though fundamental in categorization processes, is less pertinent for semantics because, as we have already seen, the most characteristic exemplar certainly does not represent the meaning of the category (sparrow is not the meaning of bird). Semantic theory's interest in psychological studies of categorization lies principally in their empirical verification of the psychological reality of the concept of regu­larity.

The idea of an underlying regularity is also present in the concept of the encyclopedia formulated by Eco, as can clearly be seen in his discussion of the concept of background and the assumptions of background proposed by Searle (1978). According to Searle, the dependence of meaning on contexts is not entirely predictable and consequently literal meaning must always be measured by a set of background assumptions that are never entirely definable and determinable, in that the specification of one throws into doubt the definition of the others, setting up a process of infinite regression. Eco (1984) contests this process of infinite re­gression and argues that there are intertextual frames or stereotypical situations establishing what we can define as the canonical form of certain contexts.

This canonical form is what I call regularity of meanings, and only in rela­tion to this regularity can one define variation and thus also understand limit situations like those described by Searle. Indeed, the paradoxical nature of the ex­amples suggested by Searle, and the immediate recognition of this (the two-meter hamburger, the hamburger placed in a block of Plexiglas, etc.) sound like an in­direct confirmation of the concept of the regularity that it seeks to deny. It is no accident, in fact, that in order to sustain the indefiniteness of meaning, Searle tesorts to the construction of deviant examples, as Cosenza (1992: 121) has ob­served. In reality, normality and variation from the norm are concepts that refer to and define each other, and it would not be possible to classify a context or meaning as deviant if there were not a canonical form from which it varied.

One must not think, therefore, that regularity excludes variation, but see these terms as the extremes of a dialectic polarity inscribed in the nature of lan­guage itself. There is a constant tension in language between stability and insta­bility, between regularity and innovation: language is neither a completely stable system of representation nor an agglomeration of continual variations without structure. Meanings are transformed not only diachronically but also synchronics ally, in that a certain degree of plasticity inheres in language, permitting ongoing adaptation to contextual variations; at the same time, there is an element of sta­bility to meaning which permits us to mean and to communicate. Language is sufficiently flexible to allow us, in an appropriate context, to describe a pile ol books as a chair, but ai the same time it is sufficiently constituted to ensure the

183

stability of the typical meaning of chair; the innovative and creative possibilities of language do not imply the absence of a stable dimension in language. On the contrary, they presuppose it.

I will now deal with the regularity with which linguistic forms are associated to typical situations, because the forms of this regularity, and their possible rep­resentations, constitute the subject matter of linguistic semantics. The role played by the stable, conventional, even repetitive aspect of linguistic functioning pro­vides a new approach ro one of the crucial problems for every discussion of mean­ing, that is, the problem of context.