Sink positive: Linguistic experience with th substitutions influences nonnative word recognition

Abstract

We used eyetracking, perceptual discrimination, and production tasks to examine the influences of perceptual similarity and linguistic experience on word recognition in nonnative (L2) speech. Eye movements to printed words were tracked while German and Dutch learners of English heard words containing one of three pronunciation variants (/t/, /s/, or /f/) of the interdental fricative /θ/. Irrespective of whether the speaker was Dutch or German, looking preferences for target words with /θ/ matched the preferences for producing /s/ variants in German speakers and /t/ variants in Dutch speakers (as determined via the production task), while a control group of English participants showed no such preferences. The perceptually most similar and most confusable /f/ variant (as determined via the discrimination task) was never preferred as a match for /θ/. These results suggest that linguistic experience with L2 pronunciations facilitates recognition of variants in an L2, with effects of frequency outweighing effects of perceptual similarity.

The influence of gradient foreign accentedness and listener experience on word recognition

“…Thus, no overt, decision-based response was required. This type of design has previously been used in a similar study investigating accented speech (Hanulíková & Weber, 2012). In the current study, the main manipulation was the degree of accentedness of the auditory items.…”

“…As the auditory stimulus unfolds and more information becomes available to the listener, the set of candidates decreases allowing listeners to settle on the intended target. However, Hanulíková and Weber (2012) have shown that this paradigm can be used to investigate the effect of accent on the time-course of word recognition as measured by looks to the target over time. In this same vein, Experiment 2 utilized the Visual World Paradigm to examine the influence of gradient foreign accentedness on the time-course of spoken word recognition.…”

“…From the work mentioned above (e.g., Dahan et al, 2008;Hanulíková & Weber, 2012;Reinisch & Weber, 2012;Witteman et al, 2013), as well as from other previous work (Sumner & Samuel, 2009;Weber, Di Betta, & McQueen, 2014;White, Yee, Blumstein, & Morgan, 2013), there is indication that listener experience (either short-term or long-term), may play a role in the ability to process accented speech. Apart from short-term adaptation, it appears that listeners also make use of long-term linguistic experience with accents to facilitate processing.…”

“…Two studies have employed the visual world/word eye-tracking methodology to investigate the recognition of non-native pronunciation variants (Hanulíková & Weber, 2012;Reinisch & Weber, 2012). First, Reinisch and Weber (2012) examined the effect of suprasegmental mispronunciation in Hungarian-accented disyllabic Dutch words.…”

“…This was interpreted as short-term, speakerspecific adaptation to suprasegmental patterns in foreign-accented speech, which aided listeners in word recognition. Hanulíková and Weber (2012) presented L2 listeners with segment-based accent variants; that is to say, the presence of the accent was signaled by changing a single segment (phonological substitution). Native English, Dutch, and German speaking participants were presented English words with three TH (/θ/) pronunciation variants, /f/, /t/, and /s/, each corresponding to typical realizations produced by English, Dutch and German speakers, respectively.…”