Planning the Articulation of Spoken Utterances

I address major controversies relating to the basic underpinnings of the human speech production system. These concern a) the nature of the mental representations of speech sounds (symbolic vs. spatiotemporal; phonemic vs. non-phonemic gestural), b) the goals of movement (articulatory vs. explicitly acoustic; with emergent vs. explicit surface durations), and c) the processes for planning and generating speech articulations (oscillator-based mechanisms vs. Optimal Control Theory and Lee’s Tau theory). PlanArt will offer the first systematic attempt to determine which theoretical assumptions about each of these three fundamental components are required to provide an adequate account of human speech production behaviour. This ambitious task requires a cutting-edge two-pronged approach. First, through novel experiments I will test predictions of theoretical alternatives for types of phonological representations, output goals, and planning processes. Many of these experiments are inspired by theory and techniques in the non-speech motor literature, which are new to the field of speech production. In the second prong of my research, I will provide a comparison of the articulatory outputs of computational models based on particular sets of core theoretical assumptions. This second, modelling prong will include the development of a flexible, modular platform allowing for parallel implementation, testing, and comparison of competing theoretical assumptions in terms of model outputs, as well as the development of a computational implementation of a new approach, proposed by team members Turk & Shattuck-Hufnagel (2020a) to address shortcomings in the literature. Taken together, this work will provide converging evidence for the nature of the bedrock elements and processes of speech production and will deliver answers to longstanding questions in phonology and phonetics about the way we plan the pronunciations of words in different contexts.