Putting prototypicality to the test

By recognising fuzzy boundaries and degrees of membership, the notion of prototypicality has made it possible to account for phenomena for which the classic, Platonic view offered no satisfactory explanation. While it originated in the field of psychology, the notion was enthusiastically adopted by cognitive linguists, who first used it with respect to the meaning of words (cf. the classical example of ‘bachelor’) and later extended it to more abstract levels of linguistic representations such as transitivity (Hopper & Thompson 1980) or past tense (Taylor 1989). The flexibility it allowed for made it an ideal tool for describing the variability of language. However, as rightly pointed out by Tsohatzidis (1990: 8), the ‘undeniable heuristic value of the notion of prototypicality should not obscure the fact that its exact theoretical shape is less clear than one might have wished, especially when it is transferred from purely psychological to specifically linguistic domains of investigation’. More specifically, it can be argued that, while cognitive linguistics has taken over the concept of prototypicality, it has failed, as a rule, to take over the methodological rigour that goes with it. Thus, many of the descriptions of prototypes found in the cognitive literature have no empirical foundation and seem to rely on the linguist’s own intuition. Besides, the term ‘prototype’ has become a ‘catch-all notion’ (Wierzbicka 1985: 343) and linguists do not always make it clear what definition they have in mind. Here, I will evoke some of the problems linked to the use of the notion of prototypicality in linguistics, such as the difficulty to test the ‘goodness-of-example’ of abstract linguistic categories, the often inevitable mediacy of language in such tests (and the possible priming effects that result from it) or the illusion that linguistic frequency can be used as a shortcut to cognitive salience. I will describe some experimental tests that were carried out in an attempt to determine the prototypical senses of highly polysemous verbs and see how the results point towards a multi-faceted characterisation of prototypicality. It will be concluded that, while there is a good case for integrating the concept of prototypicality into linguistics, it should be done with full awareness of its potential and limitations, and in an effort to clarify the concept and provide it with a firm empirical basis