Perceptual Learning of speech sounds: A bias for 'sipper' over 'zipper'?


  • Molly Babel The University of British Columbia
  • Zoe Lawler The University of British Columbia
  • Carolyn Norton The University of British Columbia


Please refer to 'supplementary file' for official abstract.


Despite huge variability in the incoming acoustical information, listeners efficiently map speech sounds into appropriate categories. Perceptual learning has been proposed as a cognitive mechanism that can account for this acoustic variation by updating a listener’s existing phonetic categories using clues from the lexical context (Norris, McQueen & Cutler, 2003). Perceptual learning of speech is especially relevant in multicultural contexts, including urban centers across Canada. Listeners are exposed to accented speech and novel pronunciations of different interlocutors in their day-to-day lives, and must accommodate these productions in order to successfully communicate.

While perceptual learning may assist in the processing of dialect and accent differences (Kraljic, Brennan & Samuel, 2008; Crista et al., 2012) and non-canonical speech (Kraljic, Brennan & Samuel, 2008b) across the lifespan (White & Aslin, 2011; Trude et al., 2013; Witteman et al., 2013), little is known about whether some pronunciations are easier to perceptually learn than others. The present study investigates whether a bias for perceptually learning typologically more prevalent devoiced fricatives over voiced fricatives exists by exposing listeners to sentence stimuli with sentence-final voiceless [s] and voiced [z] items (e.g. ‘He couldn't handle any more dinner, but there might be room for dessert’). A lexical decision task follows the exposure phase, measuring participants recalibration of [z] and [s]. Stimuli were produced naturally by one speaker, without synthesizing the placement of the critical fricatives [z] or [s] (see Weatherholtz, 2015). We predict listeners will learn the typologically more common devoicing pattern better than the voicing pattern (e.g. [ ] will be learned better than [  ]). This project contributes to our understanding of which attributes of the speech signal facilitate perceptual learning.


Author Biographies

Molly Babel, The University of British Columbia

Associate Professor, Department of Linguistics

Zoe Lawler, The University of British Columbia

Research Assistant, Department of Linguistics

Carolyn Norton, The University of British Columbia

Research Assistant, Linguistics Department

Additional Files



How to Cite

Babel M, Lawler Z, Norton C. Perceptual Learning of speech sounds: A bias for ’sipper’ over ’zipper’?. Canadian Acoustics [Internet]. 2016 Aug. 24 [cited 2024 Jun. 20];44(3). Available from:



Proceedings of the Acoustics Week in Canada