The accent atlas: A geolocation-based assessment of non-native accent familiarity and linguistic diversity

A key to successful comprehension of natural language is the ability to adapt to its variability. Recent studies have suggested that linguistic diversity in daily speech input predicts monolingual native speakers’ ability to adapt to an unfamiliar non-native accent. However, self-reports of linguistic diversity can be prone to recall errors and bias. We propose a census-based approach to quantifying the likelihood of non-native accent exposure at both the state and zip code levels. We conducted a large-scale (N = 647) conceptual replication of Clarke and Garrett (2004) to examine monolingual English listeners’ rapid adaptation to Chinese-accented speech. Results indicate that the estimated level of accent exposure at the zip code level, but not at the state level, predicts the magnitude of accent adaptation. This suggests that daily accent exposure may vary across local environments and that the current geolocation-based measure of linguistic diversity will be a useful tool to complement self-reports. Research reported in this work was supported by NIH-NICHD grant R01HD111936.