An ERP investigation of regional and foreign accent processing

Abstract

This study used event-related potentials (ERPs) to examine whether we employ the same normalisation mechanisms when processing words spoken with a regional accent or foreign accent. Our results showed that the Phonological Mapping Negativity (PMN) following the onset of the final word of sentences spoken with an unfamiliar regional accent was greater than for those produced in the listener's own accent, whilst PMN for foreign accented speech was reduced. Foreign accents also resulted in a reduction in N400 amplitude when compared to both unfamiliar regional accents and the listener's own accent, with no significant difference found between the N400 of the regional and home accents. These results suggest that regional accent related variations are normalised at the earliest stages of spoken word recognition, requiring less top-down lexical intervention than foreign accents.

Foreign-accented speech modulates linguistic anticipatory processes

“…As we mentioned above, if listeners narrow the possible lexical candidates during foreign-accented speech comprehension (Goslin et al, 2012), then we would expect that only sentences' best completions are anticipated when listening to foreign-accented speakers. We will address this issue following the rationale put forward in Federmeier and Kutas (1999) study.…”

“…This would reveal that sentences' best completions are anticipated independently of indexical properties of speakers, such as a native/ foreign accent. On the other hand, since foreign-accented speech modulates the availability of lexical and semantic information during sentence comprehension (Goslin et al, 2012;Romero-Rivas et al, 2015), it could be that anticipatory processes differ between native and foreign-accented speech comprehension. If listeners rely to a higher extent on top-down cues in order to normalize the speech signal during foreign-accented lexical processing (Goslin et al, 2012), we would expect narrowed anticipation effects during foreign-accented compared to native speech comprehension.…”

“…They observed that the final words of sentences produced by foreign-accented speakers elicited smaller N400 amplitudes compared to the native speech condition. Goslin et al (2012) concluded that foreign-accented speech could not be fully normalized during pre-lexical processing levels, therefore influencing later steps of language processing such as lexical access and integration. In other words, listeners would rely on topdown contextual cues in order to normalize the signal during foreign-accented lexical processing, narrowing the possible lexical candidates (Goslin et al, 2012).…”

“…Goslin et al (2012) concluded that foreign-accented speech could not be fully normalized during pre-lexical processing levels, therefore influencing later steps of language processing such as lexical access and integration. In other words, listeners would rely on topdown contextual cues in order to normalize the signal during foreign-accented lexical processing, narrowing the possible lexical candidates (Goslin et al, 2012). Romero-Rivas et al (2015) also explored sentence comprehension during foreign-accented speech comprehension.…”

“…Nevertheless, it is important to note that there is already some evidence suggesting that lexical-semantic processing differs during native and foreign-accented speech comprehension. Goslin et al Floccia (2012) explored lexical processing during native and foreign-accented speech comprehension, by presenting listeners with low-cloze probability 1 sentences ("His wife managed to win a bag of goldfish"). They observed that the final words of sentences produced by foreign-accented speakers elicited smaller N400 amplitudes compared to the native speech condition.…”