Date of Completion


Embargo Period



James S. Magnuson, Emily B. Myers, Gerry T. M. Altmann, Rachel M. Theodore

Field of Study

Psychological Sciences


Master of Science

Open Access

Open Access


Individual talkers vary considerably in how they produce different speech sounds, and a challenge for the listener is to learn the appropriate mapping between acoustics and phonetic categories for an individual talker. Several studies have shown that listeners are able to leverage various sources of context (e.g., coincident visual information, lexical knowledge) to guide this process, sometimes termed perceptual learning of speech. Here, we examine how sentence-level semantic information – specifically, whether preceding sentence context is predictive of an upcoming word – might modulate the size of learning effects. Across a series of perceptual learning experiments, we manipulate how learning compares between groups who receive neutral or predictive sentence contexts, also varying whether contexts are presented in the auditory or written modality. Though we observed greater learning for subjects who read predictive contexts than for subjects who read neutral contexts, this finding did not replicate in an identical follow-up experiment, suggesting that potential influences of sentence context on phonetic recalibration may be small. These findings are discussed in the context of the broader literature on perceptual learning.

Major Advisor

James Magnuson & Emily Myers