Articulation and adaptation to altered auditory feedback

Abstract

Speakers listen to themselves while talking, and they use this auditory feedback to modify their speaking plans on-line [Houde and Jordan, Science 20;279(5354):1213-1216 (1998)]. This altered auditory feedback experiment uses ultrasound tongue imaging to investigate individual differences in adaptation under three conditions: (1) raising F1 in /ɛ/, (2) raising F2 in /ʊ/, (3) raising F3 in /r/. Pilot data suggest that speakers may change both F1 and F2 in response to an altered F1, replicating Katseff et al. (2010, JASA 127(3), 1955). Principal components analysis of the ultrasound data reveals that these two acoustic changes are independently controlled. We will also test the role of individual differences in vocal tract morphology. Hard palate curvature affects variability in both articulation and acoustics [Brunner et al. (2009, JASA 125(6), 3936-3949)]: flatter palates have less acoustic stability (and greater flexibility) [Bakst & Johnson (2016, JASA 140(4), 3223-3223)], requiring greater articulatory precision to maintain acoustic consistency. We hypothesize that people with flatter palates will adapt to altered feedback faster and more completely because they (a) may have more detailed knowledge of their articulation-acoustics mapping and (b) have greater flexibility in their acoustic output.