Title: Within-syllable dynamics of vowel acoustics
Legend: We investigated whether there are differences in vowel acoustic variability between first‐ and second‐language production. Here, repeated productions of a single vowel are compared in acoustic space between a representative participant producing the French vowel [oe] as in “oeuf” (left) and the English vowel [ɛ] as in “eff” (right panel). Each open circle and connected arrowhead indicates the acoustics of a single utterance; only the most variable 33% of utterances (farthest from the acoustic median) are shown. The open circle indicates the utterance at onset (first 50 ms), while the connected arrowhead indicates the same utterance at the vowel midpoint. The spread of these utterances at onset is greater for the French vowel, at left, but both vowels show an inward movement from onset to midpoint, reducing the acoustic variability within the syllable. This tendency for greater onset variability but similar inward corrective vowel movement held over our sample of nine French learners.
Citation: Bakst, S. and Niziolek, C.A. (accepted). Self‐correction in L1 and L2 vowel production. Proceedings of the 19th International Congress of Phonetic Sciences, Melbourne, Australia, 2019.
Abstract: We listen to ourselves while talking, comparing our acoustic output to an internal auditory representation of speech targets. Previous work has shown that speakers are sensitive to their own natural acoustic variability in their native language, steering deviant productions towards auditory targets while speaking (Niziolek et al. 2013). This corrective behavior is evident in the magnitude and direction of vowel formant trajectories over the course of an utterance. In a language learned in adulthood, auditory targets may be weaker, resulting in less successful self‐correction in novel vowel categories. In the current study, participants were recorded producing monosyllabic words in L1 (English) and L2 (French). Speakers’ L2 productions showed increased acoustic variability and reduced corrective behavior compared with L1. These results are consistent with weakened auditory representations of speech targets in L2, which may impair the ability to correct one’s own productions on‐line.
About the Lab: The Niziolek lab studies how the feedback system functions in persons with aphasia, many of whom have deficits in speech production and error awareness. They are developing laboratory studies of speech skill learning that can be translated to training interventions for speakers such as these who have trouble consistently producing auditory targets. The lab’s proposed training consists of vocal games that map speech to a real-time visual display. These visual speech training games have the potential to be adapted into tools to improve speech production in individuals with speech impairment, including deaf speakers and children with developmental speech and language disorders.