Spanish is better than English for discriminating Portuguese vowels: acoustic similarity versus vowel inventory size.

Jaydene Elvin,Polina Vasiliev,Paola Escudero

doi:10.3389/fpsyg.2014.01188

Jaydene Elvin, Polina Vasiliev + Show 1 more

Open Access

https://doi.org/10.3389/fpsyg.2014.01188

Copy DOI

Abstract

Second language (L2) learners often struggle to distinguish sound contrasts that are not present in their native language (L1). Models of non-native and L2 sound perception claim that perceptual similarity between L1 and L2 sound contrasts correctly predicts discrimination by naïve listeners and L2 learners. The present study tested the explanatory power of vowel inventory size versus acoustic properties as predictors of discrimination accuracy when naïve Australian English (AusE) and Iberian Spanish (IS) listeners are presented with six Brazilian Portuguese (BP) vowel contrasts. Our results show that IS listeners outperformed AusE listeners, confirming that cross-linguistic acoustic properties, rather than cross-linguistic vowel inventory sizes, successfully predict non-native discrimination difficulty. Furthermore, acoustic distance between BP vowels and closest L1 vowels successfully predicted differential levels of difficulty among the six BP contrasts, with BP /e-i/ and /o-u/ being the most difficult for both listener groups. We discuss the importance of our findings for the adequacy of models of L2 speech perception.

Highlights

It is widely recognized that second language (L2) learners are often unable to distinguish sound contrasts that are not present in their native language (L1)
Following predictions based on vowel inventory sizes, Australian English (AusE) listeners, whose native vowel system includes all of the Brazilian Portuguese (BP) contrasts, should outperform Iberian Spanish (IS) listeners, who only have five native vowels and lack two of the mid-vowels (/E/ and /O/) present in BP, which should result in single-category assimilation and poor discrimination
The findings are in line with the L2LP model’s acoustic hypothesis and the corresponding predictions based on the detailed comparison of BP, IS, and AusE vowels that was presented in the Introduction

Summary

Introduction

It is widely recognized that second language (L2) learners are often unable to distinguish sound contrasts that are not present in their native language (L1). Models of non-native and L2 sound perception, such as the Second-Language Linguistic Perception Model (L2LP; Escudero, 2005, 2006, 2009) and the Perceptual Assimilation Model (PAM; Best, 1995) and its extension to L2 acquisition (PAM-L2; Best and Tyler, 2007) claim that perceptual similarity between native sounds and target language contrasts predicts how accurately naïve listeners and L2 learners will identify the members of those contrasts Both L2LP and PAM predict high difficulty in discrimination of target language contrasts that do not exist in the listener’s L1, which is commonly the case when the L1 has a smaller sound inventory than the L2. Escudero and Boersma (2002) suggest that multiple category assimilation may be problematic when it leads to a subset problem where the learner needs to realize on the basis of positive evidence alone that some features or vowels of their own language do not exist in the target language and may find it difficult not to perceive the extra L1 category

Methods

Results

Conclusion