Introduction

When discussing its place in language, Bolinger (1978a) characterized intonation as ‘a half-tamed savage’. This raises the question whether we can describe intonation in terms of the same structures with which we analyse the segmental morphology and phonology. We will approach this issue by considering the extent to which a number of design features of language that were proposed by Charles Hockett apply to intonation. We will conclude that intonation is indeed a half-tamed savage, but will draw a sharp dividing line between the tamed half and the untamed half. The difference lies in the discretely represented prosodic structure on the one hand, and the unusually generous scope that speakers have – compared to the latitude they have in the realization of the segmental phonology – in the phonetic implementation of fundamental frequency on the other. A rich communicative world appears to exist in this phonetic implementation component.

Broadly speaking, speakers gain control over aspects of speech production which might at first sight seem automatic, like the gradually declining pitch through the utterance, and use these for communicative purposes. In chapter 5, we will give an account of these paralinguistic meanings. At a further stage, this gradient behaviour may become grammaticalized, that is, be incorporated in the discrete structure of the language's morphology and phonology. For example, when we are more than usually concerned to get our message across, we will expend more physical effort on pronouncing the word or words that express it.