Articulatory reuse in “good-enough” speech production strategies

Abstract

Given a novel speech motor task, a speaker may optimize execution of the task for accuracy by taking feedback into account, or revert to more habitually used productions which provide a precise output not entirely optimized for accuracy. An articulatory examination comparing L1 and L2 productions was carried out in part to assess the roles of optimization and reversion to habit for individual speakers. Native speakers of American English learning French, (n = 30, 9 males) with a variety of levels of exposure to the L2, were recorded producing the monophthongs of both languages and two English approximants (/r/ and /l/) using ultrasound tongue imaging, video of lip shape, and audio. Principal component analyses run on tongue ultrasound and lip shapes of individual speakers reveal that production of L2 French essentially within L1 English articulatory habit is typical; in several cases, the approximants /r/ and /l/ are essentially reused as French vowels (e.g., /ø/ and /u/). However, a slight optimization toward native-like productions can be observed in speakers with longer exposure to the L2.