Communicatively driven versus prosodically driven hyper-articulation in Korean

Abstract

This study investigated how three different kinds of hyper-articulation, one communicatively driven (in clear speech), and two prosodically driven (with boundary and prominence/focus), are acoustic-phonetically realized in Korean. Several important points emerged from the results obtained from an acoustic study with eight speakers of Seoul Korean. First, clear speech gave rise to global modification of the temporal and prosodic structures over the course of the utterance, showing slowing down of the utterance and more prosodic phrases. Second, although the three kinds of hyper-articulation were similar in some aspects, they also differed in many aspects, suggesting that different sources of hyper-articulation are encoded separately in speech production. Third, the three kinds of hyper-articulation interacted with each other; the communicatively driven hyper-articulation was prosodically modulated, such that in a clear speech mode not every segment was hyper-articulated to the same degree, but prosodically important landmarks (e.g., in IP-initial and/or focused conditions) were weighted more. Finally, Korean, a language without lexical stress and pitch accent, showed different hyper-articulation patterns compared to other, Indo-European languages such as English—i.e., it showed more robust domain-initial strengthening effects (extended beyond the first initial segment), focus effects (extended to V1 and V2 of the entire bisyllabic test word) and no use of global F0 features in clear speech. Overall, the present study suggests that the communicatively driven and the prosodically driven hyper-articulations are intricately intertwined in ways that reflect not only interactions of principles of gestural economy and contrast enhancement, but also language-specific prosodic systems, which further modulate how the three kinds of hyper-articulations are phonetically expressed.