A task dynamic approach to prosodic strength relations (TARE)

• Applicant: Dr. Simon Roessig
• Subject Area: Applied Linguistics, Computational Linguistics
• Term: from 2022 to 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 495931197

Prosody has been shown to affect the articulatory movements of vowels and consonants systematically. At phrase boundaries and under accentuation, longer and larger movements of the tongue and the jaw can be attested. These results indicate that prosody is not only characterized by tonal aspects, e.g., intonation contours, but significantly shaped by a multitude of phonetic cues. Dynamical modelling approaches concerned with the effects on articulation have started to emerge using the concept of prosodic modulation gestures. Modulation gestures do not have constriction targets but affect the constriction gestures for the production of vowels and consonants during the mutual time of activation. While most of the models using modulation gestures concentrate on the effects of prosodic boundaries on articulation, a lot of work remains to be done with regard to the modelling of prosodic prominence. The project fills this gap in developing a dynamical model of phrase-level prominence effects on articulation using modulation gestures, based on experimental investigations. It aims at providing an adequate characterization of modulation gestures for prominence. While most studies on modulations at prosodic boundaries assume temporal modifications, e.g., slower movements, prominence-induced effects on articulation have also been found to be spatial, e.g., larger movements. The project addresses the question as to whether the spatial effects can be described as by-products of temporal modifications or whether they constitute a separate dimension of prosodic modulation. Moreover, studies on intonation indicate that the relation between multiple prominences within one phrase plays a major role. Important pragmatic meanings are not exclusively encoded locally but seem to be distributed over the whole phrase. The project therefore concentrates on the modelling of prosodic strength relations between multiple elements in one phrase and the gradiency of prosodic modulations. It uses the approach of coupled oscillations to describe effects in hierarchically organized linguistic structures. The dynamical modelling approach does not only allow for the interaction between multiple elements in the hierarchy but also captures gradual modifications. In consequence, it is able to model prosodic effects beyond the dichotomy of accented vs. unaccented. A further aim of the project is to investigate the interplay of articulatory and tonal factors. It will focus on how prosodic modulations affect supra-laryngeal articulation gestures (for consonants and vowels) and tonal gestures (for pitch accents) at the same time.

• DFG Programme: WBP Fellowship
• International Connection: USA
• Host: Professor Sam Tilsen, Ph.D.