The expectation mismatch effect in accentedness perception of Asian and Caucasian non-native speakers of English

Abstract

Previous research on speech perception has found an effect of ethnicity, such that the same audio clip may be rated more accented when presented with an Asian face (Rubin, Donald L. 1992. Nonlanguage factors affecting undergraduates’ judgments of nonnative English-speaking teaching assistants. Research in Higher Education 33(4). 511–531. doi: 10.1007/bf00973770). However, most previous work has concentrated on Asian non-native English speakers, and Caucasian speakers remain under-explored. In this study, listeners carried out an accentedness rating task using stimuli from first language Korean, German, and English speakers in 3 conditions: audio only, video only, and audiovisual. Korean speakers received similar accentedness ratings regardless of condition, but German speakers were rated significantly less accented in the video condition and more accented in the audiovisual condition than the audio one. This result is explained as an expectation mismatch effect, whereby, when the listeners saw a Caucasian speaker they did not expect to hear a foreign accent, but if they actually heard one it was made more salient by their expectation to the contrary.

The role of style in the ethnolect: Style-shifting in the use of ethnolectal features in first- and second-generation speakers

“…No effect of style found for the monolingual speakers helps us to be confident that the attested variation is not a Type 2 variation reflecting what is already present in the majority community. This lends additional support for the existence of ethnolinguistic repertoires (Benor, 2010; Clyne, et al, 2001) and Type 3 variation (Gnevsheva, 2016; Nance et al, 2016) in bilingual speakers, which may not be found in monolingual speakers of the majority variety.…”

“…when speaking about their family as opposed to their studies), which was explained through an application of the audience design and identity construction accounts (Gnevsheva, 2015). This type of sociolinguistic variation has been called Type 3 variation, expanding the model of L2 speaker variation that includes only Type 1 (variation in the acquisition of linguistic competence) and Type 2 (acquisition of sociolinguistic variation found in the L2) (Gnevsheva, 2016; Nance, McLeod, O’Rourke, and Dunmore, (2016) and Rampton (2011) also advocate for the application of the concept of style in L2). The dichotomy of Type 1 and Type 2 variation theoretically limits L2 speakers, fails to recognize their agency in the use of multiple languages, and regards any deviation as reflecting L2 acquisition failure.…”

Complicating categories: Personae mediate racialized expectations of non‐native speech

“…Prior work has examined how ratings of “accentedness” are influenced by a racialized photographic prime (e.g. Gnevsheva, ; Kang & Rubin, ; Yi et al., ; Zheng & Samuel, ). Thus, in addition to the EFA, I performed a post‐hoc analysis on the influence of voice, photograph, and their interaction on the four voice‐related rating scales implicated in the “unaccented” dimension of Factor 2 (means shown in Figure ).…”

“…However, he shows that this link enhances processing of L2 accented English when the expectation is fulfilled. Gnevsheva () provides further evidence for the “expectation mismatch” effect, finding that a mis‐match between visually expected degree of accentedness and auditorily assessed accentedness led to an even higher rating of accentedness in an audio‐visual condition.…”

Deliberative control in audiovisual sociolinguistic perception*

“…They also found that including video exaggerated the difference in perceived accentedness between the two. Gnevsheva (2018) showed that Korean L1 Korean talkers were perceived by English‐speaking listeners as accented based on audio, video and their combination, while white L1 German talkers were perceived as least accented based on video only, more based on audio and the most when in combination.…”