Towards a ‘bucket’ theory of language disability: Taking account of interaction between linguistic levels

Abstract

Models of language which recognize a series of levels are commonplace in clinical language studies, but less attention has been paid to how they interact, and the effect these interactions have on the linguistic performance of language handicapped people. Several relevant studies which seem to demonstrate the existence of restricted linguistic processing capacity are reviewed, from which it is concluded that there are a number of types of interaction which have not been sufficiently considered. A sample of spontaneous speech from a language-delayed child of 4;7 illustrates the effects of four of these interactions (syntax/non-segmental phonology, segmental phonology/other levels, syntax/semantics, pragmatics/other levels). It is concluded that there is a need for detailed description of the error-patterns of language-handicapped people, with particular reference to prosodic factors (especially those entering into the definition of fluency) to determine which kinds of interaction are most in evidence. A model of limited linguistic processing which is based on a notion of hierarchy is premature. Rather, a simpler model, referred to here as the ‘bucket’ model, in which the different levels exercise mutual influence without priorities, is more appropriate in our present state of knowledge.