UC Santa CruzLinguistics
HomeAbout the DepartmentEvents and NewsGraduate ProgramUndergraduate ProgramsPeopleResearchCourse Offerings
UCSC A-Z Index | Find People

Linguistics Dept.
UC Santa Cruz
1156 High Street
Santa Cruz, CA 95064-1077

--------------

News
Department Highlights
WHASC

--------------

Colloquia
Upcoming colloquia
Previous colloquia

--------------

Events
Current reading groups
Conferences
Other events
Non-UCSC events

--------------

Print

 


Michael Proctor

Effects of linguistic structure and bias on the use of phonetic cues in categorization

Grant McGuire
UCSC


ABSTRACT

While phonological objects are highly categorical and abstract, the phonetic facts that seem to govern them are continuous and highly variable. Therefore, it is a goal of current linguistic research to understand the process of categorization, especially the proper learning and weighting of the phonetic properties that are vital to category formation. In this talk I present results from four studies examining the shaping of phonetic categories which together demonstrate that attention to and weighting of the proper cues is directly tied to linguistic structure as well as certain biases that seem to be inherent. The first study demonstrates that a language's inventory and a bias towards formant transitions affect the cues that listeners use to differentiate place contrasts. This is followed by a second study which explores cue learning and how even brief experience facilitates category identification by warping the perceptual space in favor of the most relevant cue over a previously biased one. A third study demonstrates that the distribution of cues in a multidimensional space affects infant acquisition of categories and that a specific bias towards transition information is not operative during early acquisition; either speech production or phonological structure must drive this effect. A final study on generalization and integration of cues demonstrates that cues can be generalized and abstracted to new contexts, but that the transition bias may be a more general phenomenon that is not necessarily rooted in formant movement, but more generally in attention to periodic information.