Introduction

The assumption of phonological structure is so deeply embedded in instrumental phonetics that it is easy to overlook the ways in which it directs our investigations. Imagine a study of “acoustic cues to negation” in which it is concluded, by a comparison of spectrographic analyses of negative and corresponding affirmative utterances, that the occurrence of nasalized formants shows a statistically significant association with the expression of negation in many European languages. It is quite conceivable that such data could be extracted from an instrumental study, but it is most unlikely that anyone's interpretation of such a study would resemble the summary statement just given. Nasalized formants are acoustic cues to nasal segments – such as those that happen to occur in negative words like not, nothing, never (or non, niente, mai or n'e, n'ikto, n'ikogda, etc.) – rather than direct signals of meanings like “negation.” The relevance of a phonological level of description – an abstraction that mediates between meaningful units and acoustic/articulatory parameters – is taken for granted in any useful interpretation of instrumental findings.

Until recently, the same sort of abstraction has been all but absent from instrumental work on intonation. Studies directly analogous to the hypothetical example have dominated the experimental literature, and the expression “intonational phonology” is likely to strike many readers as perverse or contradictory.