Acoustic Effects of Prosodic Boundary on Vowels in American English

0. Introduction Prosody in speech is expressed through the modulation of pitch, duration, loudness and spectral balance. These features encode (i) phonological phrase structures that serve to group words into sense units, and (ii) prominence structures that mark the location of perceptually prominent syllables and words within larger domains. The prosodic phonological structures that encode phrasing and prominence are hierarchically organized: smaller/lower phrases are contained within larger/higher ones, and prominent units defined within a larger domain are projections of prominent units defined in successively lower domains. The hierarchical model of prosodic structure posits that the distinction between smaller and larger prosodic domains is a distinction in level. The effects of prosodic domains on phonological and phonetic structures are similar in type across levels, but levels may differ in the number or strength of those effects. This paper investigates the presence of acoustic correlates of the level distinction between prosodic domains. The focus is on the distinction between the smaller 'intermediate' phrase and the larger 'intonational' phrase through a comparison of the effects of the phrasal boundary on the final vowel in the phrase. The study reported here addresses the question of whether the effects of the higher phrase boundary on acoustic measures are of the same type as the effects of the lower boundary, and if so, whether the higher phrase level exhibits acoustic effects with consistently stronger magnitude compared to the lower phrase level. The goal is to discover if there is evidence that prosodic phrase boundaries produce acoustic effects that are cumulative over successive layers of prosodic structure. Earlier research has found that articulatory speech gestures undergo spatiotemporal enhancement when they occur at the left or right edge of a prosodic In domain-edge positions, the speech gestures that form constrictions in the vocal tract are produced with temporally longer duration, greater linguopalatal contact and larger displacement. These effects are modeled by Byrd (2000) and Byrd and Saltzman (2003) in terms of a prosodic " π-gesture " , which warps the temporal dimension of speech in the region of a juncture