Presentation

Cristal Giorio Jackson¹, Janet G. van Hell¹; ¹Pennsylvania State University

In the current globalized world, listeners are exposed to different types of accented speech. Interestingly, behavioral and electrophysiological research has shown that listeners who speak in the same nonnative accent as the talker find them equally or even more intelligible than a native speaker of that language, known as the Interlanguage Speech Intelligibility (ISI) Benefit (Bent & Bradlow, 2003). ISI behavioral studies have found that nonnative listeners rate other nonnative talkers with a shared first language (L1) as more intelligible (Hayes-Harb et al., 2008; Stibbard & Lee, 2006). Event-Related Potential (ERP) studies testing nonnative-accented speech have found that processing semantically anomalous sentences spoken by a nonnative talker entails a processing cost shown by a delayed N400 effect (Goslin et al., 2012; Grey & van Hell, 2017). These studies typically treat a given nonnative accent as a collective accent shared across all talkers, such as Spanish-accented English in the Spanish speaking community. However, a given language (e.g., Spanish) has accent variants and each variant has distinct and recognizable sound features; for example, over 20 countries have Spanish as their official language, resulting in various distinct Spanish accent variants. Sound features of an L1 impact the production and acoustic-phonetic characteristics of the later-learned second language (L2), and different L1 variants produce distinct L2 nonnative accents (Bent & Frush Holt, 2013). This ERP study examines: 1) How Spanish-English bilingual listeners process Spanish-accented English sentences produced by talkers who speak their own variety of Spanish (i.e., Mexican-Spanish) or a different variety of Spanish (i.e., Chilean-Spanish); and 2) If the ISI benefit effect extends to processing different varieties of Spanish. In this study, 35 highly proficient Mexico Spanish-English bilinguals will listen to semantically anomalous and correct sentences spoken in English by bilingual talkers with a matched accent variant (Mexico Spanish-accented English) and a mismatch variant (Chile Spanish-accented English). Additionally, sentences will be spoken by a control condition of native-accented English (i.e., American-English), and a control condition of nonnative-accented English (i.e., Chinese-accented English) speakers. We hypothesize that if the ISI benefit is restricted to listeners’ own variant of Spanish, then semantic anomalies in the accent-variant that matches the listeners’ accent are predicted to elicit a canonical N400 response, which is also predicted for the native-accented English control condition. Semantic anomalies in the mismatched Spanish variant are predicted to elicit a delayed N400, as is also predicted for the nonnative-accented English control condition. In contrast, if the ISI benefit extends across all accent variants of a given language (here: Spanish), then semantic anomalies in both Mexican- and Chilean-Spanish accented English sentences will elicit a canonical N400 in Mexican-Spanish-English bilingual listeners, similar to that observed in the English native-accented sentences, but markedly different from the delayed N400 response to the nonnative-accented English control sentences. Critically, the latter result pattern would reflect a theoretically significant broadened scope of the ISI benefit: not only the listeners’ own nonnative accent but also a variant of that accent, which has distinctively different sound features, benefit listeners’ processing of nonnative accented speech.

Topic Areas: Speech Perception, Multilingualism