Published July 7, 2021 | Version v1
Conference paper Open

Relative functional load determines co-articulatory movements of the tongue tip

  • 1. University of Tübingen

Description

Does morphological structure affect articulation when segmental similarity is strictly controlled? To address this question, we used electromagnetic articulography to study the articulatory trajectories of tongue tip and tongue body during the articulation of German words containing [a(:)] as stem vowels followed by [t] that in roughly half of the words realized an inflectional function. According to a generalized additive model fitted to the articulatory trajectories of tongue body and tongue tip sensors using electromagnetic articulography, a factorial predictor signaling the presence or absence of an inflectional exponent underperformed compared to a quantitative measure derived from a Linear Discriminative Learning model. This quantitative measure is based on the functional load of triphones, i.e., the extent to which a triphone contributes to the meaning of the word. The relative functional load of the stem triphone (centered around [a(:)]) and the triphone pivoted on the [t] emerged as a strong co-determinant of articulation. Importantly, words with a balanced relative functional load (i.e., a value close to zero) revealed optimized smooth co-articulation at tongue body and tongue tip sensors. These results provide evidence for the possibility that differences in the details of articulation straightforwardly reflect differences in meaning as captured by distributional semantics.
 

Notes

This study was funded by the European Research Council (WIDE-#742545 awarded to the third author) and by the Deutsche Forschungsgemeinschaft (Research Unit FOR2373 'Spoken Morphology', Project 'The articulation of morphologically complex words', BA3080/3-2).

Files

Saito2021_ISSP2020_EMA.pdf

Files (594.5 kB)

Name Size Download all
md5:c97271ae7ad3f28fe6083f4f6d12714f
594.5 kB Preview Download

Additional details

Funding

WIDE – Wide Incremental learning with Discrimination nEtworks 742545
European Commission