Optimization of articulator trajectories in producing learned nonsense words

• Published in: The Journal of the Acoustical Society of America 2009-04, Vol.125 (4_Supplement), p.2499-2499
• Main Authors: Tiede, Mark, Mooshammer, Christine, Goldstein, Louis, Shattuck-Hufnagel, Stefanie, Perkell, Joseph, Matthies, Melanie
• Format: Article
• Language: English
• ISSN: 0001-4966; 1520-8524
• DOI: 10.1121/1.4783359

Previous work assessing motor learning in novel speech utterances has shown that kinematic duration and variability are reduced as a function of practice [Schulz et al. (2001)]. In this study EMA was used to observe articulator movements during production of a variety of polysyllabic nonsense words conforming to English phonotactics (e.g., “thraimpoframodis”). Target words were elicited eight times in block-random order, with each block separated by ten minutes of an unrelated task serving as a distractor. Analysis of EMA sensor trajectories between initial and final repetitions shows in general reduction of overall duration, lower STI [Smith et al. (1995)], reduced stiffness, and fewer acceleration peaks, with intermediate productions indicating that these trends occur asymptotically. Comparison of consonant closure timings delimited using velocity extrema and aligned through nonlinear time normalization [Lucero et al. (1997)] suggests that as fluency increases, the overlap between adjacent consonantal gestures also increases. The patterns of articulatory optimization illustrated by these results will be contrasted with similar analyses of real but difficult-to-pronounce sequences (tongue twisters) collected from the same subjects, with the goal of identifying similarities between pre-fluent and dysfluent utterances. [Research supported by NIH.]