Abstract

In recent years, much has been said about the ways in which figurative language is processed. Two of the principal trends offer apparently irreconcilable views on the subject. On the one hand, the ‘direct-access view’ (Gibbs 2002) argues that the processing of figurative language does not differ significantly from that of non-figurative language. Contrasting with this is the ‘graded salience hypothesis’ (Giora 2002), in which it is argued that the successful interpretation of a language item is determined by its degree of saliency. With regard to these standpoints it must be highlighted that each addresses a different type of figurative language: the direct-access view focuses its investigations on conventional and familiar idioms and metaphors, whereas the graded salience hypothesis concentrates on unconventional and unfamiliar expressions. It is almost certainly for this reason that they are unable to reconcile their differences, as different objects of study give rise to different results and hypotheses. This research (Philip 2003) started out as an attempt to understand when, and under what conditions, the connotative meanings of colour words became activated in idiomatic and metaphorical expressions such as ‘see red’ and ‘green with envy’. Many explanations exist to account for their figurative and connotative meanings in etymological terms, but none contribute satisfactorily to defining their contextualised meaning in a synchronic perspective. The data analysed came from two large general reference language corpora*, and produced surprising results. Insofar as canonical forms of the expressions studied are concerned, connotative meanings are not activated; in fact quite the opposite occurs. The words composing the phraseological chunks become semantically washed-out, or ‘delexicalised’ (Sinclair 1991). The chunks are not decomposed into their component parts, and appear to be processed in the same way as non-figurative and one-word members of the lexicon, thus supporting the direct-access view. Yet canonical forms of figurative expressions are quite rare. Language users are creative, and they bend and twist the conventional to suit their communicative purpose. The extremes of this are highly visible in newspaper headlines and the like, where fixed phrases provide a structure upon which deliberate puns can be coined. When conventional idioms and metaphors undergo variation the delexicalisation process is reversed: the new elements have to be incorporated into the existing, known structure, and their relation to the context ascertained so that the intended meaning can be appreciated. This process, relexicalisation, is largely dependent on the identification of salient meanings, and therefore lends credence to the graded salience hypothesis. The results of this study provide substantial evidence that the difference between the literal and the metaphorical can be more accurately defined as a difference between the conventional and the innovative. In between the extremes of canonical forms and deliberate, contextually-appropriate variation, there is a cline through which both grammatical and lexical variation progresses. It is this little-studied middle-ground that demands the attention of psycholinguists and corpus linguists alike if the linguistic processes in operation are ever to be fully understood.