Arjona Building

Shawn Cummings, in the Language & Cognition division of the Department of Psychological Sciences, will be defending his dissertation: "Linking Lexically Guided Perceptual Learning to Statistical Patterns in Speech Input". ABSTRACT: Listeners use lexical information to modify the mapping between speech acoustics and speech sound categories. Despite convention to consider lexically guided perceptual learning as a binary outcome, the magnitude of the learning effect varies in the extant literature. We hypothesize that graded learning outcomes can be linked, in part, to statistical characteristics of the to-be-learned input, consistent with the ideal adapter theory of speech adaptation. We begin with creation and analysis of a lexically guided perceptual learning corpus including stimulus sets for the /ʃ/-/s/ contrast for each of 16 talkers following standard methods (i.e., waveform averaging to create ambiguous variants), yielding variability in the statistical cues specifying this contrast across talkers (Experiment 2). We then analyze the perceptual consequences of this variability on perception prior to experimentally induced bias (Experiment 2). Finally, we (a) measure lexically guided perceptual learning for each talker, (b) identify input characteristics that are associated with learning magnitude, and (c) examine whether a computational instantiation of the ideal adapter theory can model the input-learning link (Experiment 3). Robust learning is observed for 13 of 16 talkers, with magnitudes of learning strongly convergent between behavior and model simulations. These results provide a critical and successful test of the ideal adapter framework for speech adaptation, thus informing an understanding of the mechanisms that allow listeners to solve the lack of invariance problem for speech perception.