Kraljic, T., Brennan, S. E., & Samuel, A. G. (2008). Accommodating variation: Dialects, idiolects, and speech processing. Cognition,107, 54–81.
Presentation: Sarah
Summary: Shelly
The ability of listeners to deal with enormous variation in pronunciation has been a topic attracting much attention in speech perception. However, less have been dedicated to how different sources of variation might influence such an ability. This question arises from the fact that variations from different sources are usually driven by different forces, and listeners might develop different processes to cope with these variations. For example, in the dialect of New York/Long Island English, the consonant /s/ is usually realized as [ʃ] when followed by [tr]. When facing contextually-driven variations as this, listeners could uncover the underlying /s/ by adopting the rule of co-articulation. However, this strategy may fail, if the /s/ ® [ʃ] phenomenon is an idiolectal variation and not contextually-driven, then listeners may need to restructure its phonemic representation to accommodate the new pronunciation. Therefore, if the hypothesis is correct, then one should expect that listeners should have more perceptual learning when exposed to idiolectal [ʃ] than when exposed to dialectal [ʃ]. To test this hypothesis, two experiments were conducted. In Experiment 1, subjects were divided into two groups. One was exposed to material where the /s/’s before [tr] were pronounced as [ʃ] (dialectal); the other was exposed to material where all /s/’s were pronounced as [ʃ] regardless of the context (idiolectal). Results showed that if asked to categorize the two sounds from a [asi]-[aʃi] continuum, subjects of idiolectal group would categorize significantly fewer items as [ʃ] than the other group, meaning that they have reconstructed the phonemic representation of [s], and perceptually learn to identify more [ʃ]-like sounds as [s]. In Experiment 2, this perceptual contrast was replicated. In addition, production tasks were performed before and after the perception test, to see whether such perceptual learning can be carried over to listeners’ production. However, no significant results were found in production. The current study showed that the perceptual system does not treat all variations in pronunciation equally. Variations from difference sources are treated differently even when they have identical acoustic realizations.
