Intrinsic coherence of prosodic and segmental aspects of speech

It has been long held that segmental and prosodic aspects of speech belong to very different domains. Segments refer to vowels and consonants which constitute the phonemes that make up the words of speech, and they are what most writing systems record. As such they are viewed as the core components of speech. Prosodic components, also known as suprasegmentals, refer to properties like stress, timing, duration and intonation. According to Lehiste (1996:227): “segmental features characterize speech sounds, and suprasegmental features are properties of speech sounds or their sequences that are simultaneously present, that do not change the distinctive phonetic quality of the speech sounds, but do modify the sounds in a way that may change the meaning of the utterance”. In this definition, speech sounds refer only to consonants and vowels, and suprasegmental properties are treated as properties that accompany the speech sounds. Although speech research used to be dominated by segmental studies that are concerned with only consonants and vowels, in recent years speech prosody has become an increasingly active area of research. Yet prosody is still typically viewed as not only separate from segments, but also involving rather different articulatory and perceptual mechanisms. The goal of this paper is to demonstrate that this conceptual divide is unwarranted. We will show that segments and prosody are in fact intrinsically coherent, because they share essentially the same articulatory dynamics, and are both involved in encoding lexical as well as non-lexical communicative functions. We will make our case by first reviewing evidence for the basic articulatory mechanisms of tonal dynamics, and then exploring how the same principles can be extended to segmental and intonational dynamics. Finally, we will examine how tonal as well as segmental reduction can be explained in terms of both articulatory dynamics and functional requirements.