Postdoctoral Fellowship: SPRF: Mechanisms Underlying Perceptual Learning of Accented Speech

Melguy, Yevgeniy, Berkeley CA

Investigators

Abstract

This award was provided as part of the NSF Social, Behavioral and Economic Sciences Postdoctoral Research Fellowships (SPRF) and Linguistics programs. The goal of the SPRF program is to prepare promising, early career doctoral-level scientists for scientific careers in academia, industry or private sector, and government. SPRF awards involve two years of training under the sponsorship of established scientists and encourage Postdoctoral Fellows to perform independent research. NSF seeks to promote the participation of scientists from all segments of the scientific community, including those from underrepresented groups, in its research programs and activities; the postdoctoral period is considered to be an important level of professional development in attaining this goal. Each Postdoctoral Fellow must address important scientific questions that advance their respective disciplinary fields. Under the sponsorship of Dr. Melissa Baese-Berk at the University of Chicago, this postdoctoral fellowship award supports an early career scientist investigating perceptual learning of accented speech. Listeners often have trouble understanding an unfamiliar accent or dialect, but comprehension can rapidly improve as they become more familiar with a given speaker. However, research shows that such learning is possible and feasible, it is unknown exactly which strategies listeners rely on in order to improve their comprehension. This project uses artificial accent learning tools to investigate how listeners learn to adapt to accented speech, and how in turn this knowledge may be used to benefit listeners in novel listening contexts. Through this study, we hope to shed light on how listeners can successfully adapt to the diversity of speech contexts that we encounter in the real world, rather than expecting all speakers to sound the same. Normalizing speech variability in this way can have important societal ramifications for speakers who come from diverse linguistic backgrounds. Despite several decades of research on perceptual learning for speech, the mechanisms that underlie listener adaptation to accented speech are still poorly understood. Existing research has largely addressed this question using one of two approaches. The first has focused on natural accents, whereas the second has used artificial accents, created by manipulating individual target sounds in otherwise natively-accented speech. Both approaches have issues. The first is much more representative of the real world-task listeners face, but offers limited insight into the mechanisms of perceptual learning, because natural accents differ on many different dimensions. By contrast, the second approach gives a clearer idea of how accent exposure changes perception of specific sound categories but represents a less ecologically-valid scenario. This project would attempt to bridge these parallel literatures, using well-established research methods (e.g., lexically-guided recalibration) to investigate the mechanisms at work in listener adaptation to non-natively accented speech. In particular, we hope to shed light on the role of high-variability training protocols and cross-speaker variability in achieving robust accent learning that can generalize across related listening contexts. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.