Native-accented Speech Research Articles

Foreign-accented speaker identity affects neural correlates of language comprehension

Abstract This study tested semantic and grammatical processing of native- and foreign-accented speech. Monolinguals with little experience with foreign-accented speech listened to sentences spoken by foreign-accented and native-accented speakers while their brain activity was recorded using EEG/ERPs. We gathered behavioral measures of sentence comprehension, language attitudes, and accent perception. Behavioral results showed that listeners were highly accurate in comprehending both native- and foreign-accented sentences. ERP results showed that grammatical and semantic violations elicited different neural responses in native versus foreign accented speech. Native-accented speech elicited a frontal negativity (Nref) for grammatical violations and a robust N400 for semantic violations. However, in foreign-accented speech only semantic (not grammatical) violations elicited an ERP effect, a late negativity. Closer inspection of listeners who did and who did not correctly identify the foreign accent revealed that listeners who identified the foreign accent showed ERP responses for both grammatical and semantic errors: an N400-like effect to grammatical errors and a late negativity to semantic errors. In contrast, listeners who did not correctly identify the foreign accent showed no ERP responses to grammatical errors in the foreign-accented condition, but did show a late negativity to semantic errors. These findings provide novel insights into understanding the effects of listener experience and foreign-accented speaker identity on the neural correlates of language processing.

Relative Difficulty of Understanding Foreign Accents as a Marker of Proficiency.

Foreign-accented speech is generally harder to understand than native-accented speech. This difficulty is reduced for non-native listeners who share their first language with the non-native speaker. It is currently unclear, however, how non-native listeners deal with foreign-accented speech produced by speakers of a different language. We show that the process of (second) language acquisition is associated with an increase in the relative difficulty of processing foreign-accented speech. Therefore, experiencing greater relative difficulty with foreign-accented speech compared with native speech is a marker of language proficiency. These results contribute to our understanding of how phonological categories are acquired during second language learning.

Categorization training for non-native accented word recognition

Non-native accented speech is considered an adverse listening condition for native speakers (Mattys et al., 2012). Previously, accurate categorization of non-native accents was found to relate to accented speech-in-noise recognition (Atagi and Bent, 2015). The goal of the current study is to examine (1) the extent to which native speakers can be trained to categorize the language backgrounds of non-native talkers using feedback, and (2) the extent to which accent categorization training, relative to word and sex identification tasks, improves non-native accented word recognition. Participants were randomly assigned to one of three training tasks: an accent categorization task with trial-by-trial feedback (n = 10), a transcription task (n = 10), and a speaker sex categorization task (n = 10). Critically, all participants were exposed to the same set of training stimuli. Participants completed a transcription task with speech-shaped noise prior to and following training. We assessed keyword accuracy for non-native accented speech and native accented speech. Our results demonstrate that single-session training can improve accent categorization. Accent identification training, word identification, and sex identification all yielded more accurate non-native speech perception (post>pre). We find that irrespective of training task, exposure to non-native accents enhances intelligibility of non-native accented speech.

Recognition memory for foreign- and native-accented sentences

The acoustic-phonetic properties of speech signals affect not only their intelligibility, but also how well they are encoded in memory. Recognition memory can be improved, for example, when speakers intentionally speak clearly (Van Engen et al., 2012). One explanation for this result is that enhanced acoustic-phonetic cues reduce the cognitive effort associated with perceptual speech processing, thereby increasing the availability of processing resources for encoding speech content in memory. In the current study, we tested the hypothesis that, following the same logic, recognition memory would be reduced for foreign-accented speech, in which acoustic-phonetic cues deviate from native-language norms. Participants heard English sentences produced by a native speaker of English and a native speaker of Korean. Recognition memory was tested using an old-new judgment task in which participants heard those recordings again, along with an equal number of new items. Surprisingly, recognition memory was higher for Korean-accented English than for native-accented English. A follow-up study in which test sentences were presented visually showed no difference between foreign- and native-accented speech. These results are consistent with an acoustic distinctiveness account in which foreign-accented speech can in some cases be remembered more accurately due to its distinct acoustic features.

Hot topics in speech communication: Listening to foreign-accented speech

There are currently more non-native English speakers in the world than there are native speakers. Most of these second language users will speak with a detectable foreign accent. Foreign-accented speech differs from native language norms along many acoustic-phonetic dimensions including the realization of vowel, consonant, and prosodic features. An important question for researchers in the field of speech communication is how this type of language variation influences speech perception and perceptual processing. Numerous findings have shown that foreign-accented speech is generally less intelligible, receives lower comprehensibility ratings, and is processed more slowly than native-accented speech. Further, these negative perceptual effects can be exacerbated by noisy listening conditions or listener variables such as age or hearing loss. However, research over the past several years has shown the amazing flexibility of the speech perception mechanism in its ability to adapt to this form of variability. Through experience and training, listeners can improve their word identification skills with specific foreign-accented talkers, particular foreign accents, and foreign-accented speech in general. New directions in this research area include perception of foreign-accented speech by infants and children as well as how a foreign accent may influence memory.

Examining talker effects in the perception of native- and foreign-accented speech

Understanding the circumstances under which talker (and other types of) variability affects language perception represents an important area of research in the field of spoken word recognition. Previous work has demonstrated that talker effects are more likely when processing is relatively slow (McLennan & Luce, 2005). Given that listeners may take longer to process foreign-accented speech than native-accented speech (Munro & Derwing, Language and Speech, 38, 289-306 1995), talker effects should be more likely when listeners are presented with words spoken in a foreign accent than when they are presented with those same words spoken in a native accent. The results of two experiments, conducted in two different countries and in two different languages, are consistent with this prediction.

Recognition memory for speech of varying intelligibility

Previous research has demonstrated that the internal representation of spoken words includes both a phonetic description and information about the source characteristics of specific talkers. Spoken utterances also vary greatly in intelligibility. However, relatively little is known about how such variability affects the encoding of speech signals in memory. In the current study, we hypothesize that increased intelligibility leads to better recognition memory for speech. We test this hypothesis by comparing recognition memory for clear and conversational speech produced by native and non-native speakers of English. In experiment 1, listeners were exposed to semantically anomalous sentences produced by a native speaker of English in clear and conversational speech styles. In experiment 2, listeners heard sentences in both styles produced by a Croatian-accented speaker of English. In each experiment, after initial exposure, recognition memory was tested with an old-new identification task. Preliminary results suggest that native-accented speech and clear speech may facilitate recognition memory compared to foreign-accented and conversational speech. The neural mechanisms involved in processing speech of varying intelligibility are currently being investigated using fMRI. The results of this project will expand our understanding of how different sources of variability in the speech signal affect speech processing and memory representations.