Rich Reduction: Sound-segment residuals and the encoding of communicative functions along the hypo-hyper scale

Abstract

The H&H (Hypo-Hyper) Theory of Lindblom (1990) is probably one of most prominent theories of the phonetic sciences. It has considerably shaped and advanced the picture that we have of speech reduction today, and it belongs to the standard repertoire of every introductory course in phonetics. The theory was put forward at a time when research on speech reduction began to undergo a shift in focus from the description and linguistic embedding of phonological processes like assimilation, elision and lenition to questions about their phonetic details, contextual factors, perception, and cognitive processing. The reason behind this shift was the growing relevance of connected speech, driven by the advent of powerful computer-based models (e.g., of articulation) and digital recordings of several hours of actual spontaneous dialogues. The consistent application of these digital technologies and resources - and not least the progress that was made on this basis in speech synthesis - have fundamentally changed our knowledge of speech reduction. We know today that speech reduction is not a rare phenomenon of certain speaking styles, word classes or phrases, and that it is not limited to individual sounds and manifestations. These insights challenge the scientific (not didactic!) value of established instruments like the phoneme and the canonical form. The present chapter will argue with reference to examples from different languages - and in accord with Lindblom's own expectation - that also the H&H theory needs to be adapted in the light of current knowledge. This does not concern basic concepts like gradual variation in reduction and the anticipation of listener knowledge in speech production. However, the situation is different with Lindblom's "tug-of-war" metaphor that conceptualizes the realized degree of reduction as a compromise between economic and intelligible speech. First, growing perception evidence questions the metaphor's key assumption that more articulatory economy and hence a higher degree of reduction make speech less intelligible for listeners. Moreover, a one-dimensional hypo-hyper continuum controlled by two antagonistic forces (speaker and listener) ignores that fact that communicative functions are another separate driving force for variation in the degree of reduction. Therefore, the author suggests to abandon that the tug-of-war metaphor in favor of an adaptation Bolinger's famous wave metaphor.