Adaptive targets of perturbed fricatives

The aim of the project is to characterize speakers’ sound representations by investigating their compensatory behaviour when their speech is auditorily perturbed.
Sound representations have been described as being predominantly articulatory or acoustic in nature. Arguments for articulatory sound representations have been found in diachronic sound shifts. During the High German consonant shift, for example, /t/ was shifted to /s/. Within this period there must have been a time during which speakers perceived both, /t/ and /s/ as „the same“. While, acoustically, these sounds are very different, they are similar articulatorily. /s/ has a tongue position that is just slightly lower than /t/. Therefore, what speakers led to perceive the sounds as similar must have been the articulatory properties of the sounds. This reasoning led to the conclusion that sound representations are articulatory, in other words: If speakers aim to produce a sound, what they have in mind is a certain articulatory configuration. The acoustic result is of secondary importance.
Several studies, including our own work, however, challenge this view. During the first phase of the project we investigated Russian speakers’ reaction to a perturbation of a vowel. In this experiment the speakers produced a sound that was manipulated acoustically (the 2nd formant was shifted) and fed back to the speaker via insert earphones. Depending on the context F2 was shifted upwards or downwards. The great majority of the subjects compensated: whenever the formant was shifted upwards they adapted their articulation and produced a lower second formant and vice versa. Depending on the direction of the shift the subjects used two different configurations for one and the same vowel. This result speaks for acoustic rather than articulatory sound representations.
A similar experiment was carried out with sibilants. In these sounds the center of gravity was shifted. In this second experiment, however, the subjects generally compensated in one direction only.
The aim of the project described in the renewal proposal is to investigate whether the differences in the adaptation are due to differences in the sound representations of vowels versus consonants. Alternatively, since the adaptation of sibilants is quite demanding, it is possible that the subjects had not finished the adaptation process. While in the vowel experiment the speakers just needed to change the tongue position globally, for the adaptation of the sibilants they needed to adapt at least two other parameters, i.e the tongue shape and the height of the jaw. These parameters can only be analysed with the help of articulatory data (Electromagnetic Articulography). The aim of the second project phase is therefore to investigate speakers’ articulatory strategies. Our hypothesis is that speakers will use two different articulatory strategies in order to adapt for the two perturbation directions.

Principal Investigators
Brunner, Jana Dr. phil. (Details) (German Linguistics: Phonetics / Phonology)

Financer
DFG: Sachbeihilfe

Duration of Project
Start date: 09/2018
End date: 02/2020

Research Areas
Linguistics

Research Areas
Phonetik

Publications
Klein, E., Brunner, J. & Hoole, P. (2019). The relevance of feedback for consonant production: the case of fricatives. Journal of Phonetics (77).
Klein, E., Brunner, J. & Hoole, P. (2019). Spatial and temporal variability of corrective speech movements as revealed by vowel formants during sensorimotor learning. Speech production and perception: Learning and memory. Peter Lang. p. 77-107.

Last updated on 2020-01-06 at 19:38