Catégorie d'événements : SRPP

In this talk, I will present a lip tracking method that can be used to study speech articulation and confront articulatory data with acoustic evidence of stop lenition. While earlier motion capture methods required special equipment, new developments were made in recent years that allow for optical lip tracking based on internet camera recordings. I will discuss two of my studies that use this method and implementations presented by other authors that confirm its validity for segmental research.
First, I will show that looking at lip aperture measurements helps determine the contexts and extent of /p b/ weakening in Canary Islands Spanish. Lip tracking data reveals that:

– stop lenition is limited to postvocalic position
– stop lenition Is blocked in derived postvocalic position
– there are asymmetries in prosodic effects when compared to acoustics.

The second study looks at final /t d/ deletion in American English. Here, the discrepancy between articulation and acoustics is under analysis and lip movement data is used to test the completeness of the process. /t d/ trapped in between bilabials are used to track lip aperture changes that may be semblant of incomplete pronunciations. Ultrasound data of tongue blade movements is used to confirm lip trajectories as proxies of alveolar gestures.

Segments within the domain of stress undergo articulatory strengthening along two dimensions: time, involving enhancement of gesture duration, and space, involving enhancement of gesture displacement. However, the way in which these dimensions interact with each other across phonetic contexts remains underinvestigated. Moreover, while research on stress has mainly focused on stressed vowels, recent findings suggest that the effects of stress may extend over a broader window. In this talk, we present an investigation of stress-induced articulatory strengthening in Italian, in which we jointly analyzed the temporal and spatial characteristics of articulatory gestures using a principal component framework. Our findings reveal that temporal and spatial modulations constitute independent dimensions of stress-induced strengthening, whose relative contribution varies across segments. We also show that both dimensions extend beyond the stressed vowel, affecting the realization of neighboring consonants and vowels. We discuss these results in light of articulatory models of prosody in which stress is represented as a modulation gesture scoping over prosodic constituents and shaping the spatiotemporal organization of the constriction gestures realized within its domain.

This talk presents two research projects themed on tone development and change. The first part focuses on contemporary tonal variation, through an acoustic-phonetic case study of tones in Plastic Mandarin, a newly crystallised urban variety spoken by younger generations in Changsha, China. Drawing on fieldwork speech data and analysis using Generalised Additive Mixed Models, the study reveals how citation tones and neutral tone are realised in both Standard and Plastic Mandarin and situates the findings within the framework of new dialect formation. The second part looks to the past, exploring a possible pathway of tone emergence through a psychoacoustic experiment. This on-going big team science project aims to advance our understanding of microprosodic pitch perception in speech. Together, these studies offer new insights into the mechanisms underlying tone innovation and change in human language.

This study examines the phonetic basis of a contrast described between plain and uvularized vowels in Mawo Qiang (Qiangic, Sino-Tibetan). Building on Evans et al. (2016), it adopts their analytical framework while expanding the empirical basis through newly collected data from nine speakers, analyzed using acoustic measurements and articulatory modelling. The acoustic results show that the contrast is primarily signaled by lowering of F2, whereas changes in F1 and higher-formant measures are weaker and more variable, depending on vowel quality and speaker. The articulatory modelling suggests that vowels described as uvularized often involve posterior displacement of the tongue dorsum, but this is neither the sole nor a uniform strategy. Across vowel pairs, the contrast may also involve tongue lowering, pharyngealization-like configurations, or differences in vowel fronting rather than straightforward retraction. Taken together, these findings indicate that the opposition is acoustically robust but phonetically heterogeneous. Rather than reflecting a single invariant uvular gesture, the data indicate multiple articulatory strategies converging on a similar acoustic effect, particularly F2 lowering. Accordingly, if the contrast is phonologically structured as proposed in previous work, its phonetic realization may be better understood in terms of relative fronting versus retraction rather than uniform uvular secondary articulation. These results contribute to discussions of secondary vowel articulations in Qiangic languages. They also highlight the importance of distinguishing between phonological categories and their phonetic realization, and the value of detailed instrumental investigation in refining the articulatory characterization of typologically rare contrasts.