Date of Completion

5-9-2015

Embargo Period

4-28-2017

Advisors

Carl Coelho, Adrian Garcia-Sierra

Field of Study

Speech, Language, and Hearing Sciences

Degree

Master of Arts

Open Access

Open Access

Abstract

The acoustic signal of speech cues information about who is speaking in addition to a talker’s conceptual message. Recent findings indicate that these two aspects of the acoustic signal are fundamentally intertwined in the context of speech perception. For example, listeners demonstrate a native-language advantage for talker identification, which has been interpreted as evidence that phonological knowledge is recruited for talker identification. Converging evidence for this account comes from studies indicating adults with reading disability due to deficits in phonological processing show impaired talker recognition even in their native language. Other studies suggest that the influence of phonological processing on talker identification is a gradient one, such that the detriment listeners experience when identifying talkers of non-native language reflects a continuum of experience and expertise with that language. Here we test the hypothesis that the gradient influences of reading ability on native and non-native talker identification will be observed for the range of reading abilities that mark unimpaired readers. Monolingual, English readers were assigned to either the average (n = 17) or advanced (n = 17) reading group based on a median split of aggregate performance across a standardized assessment battery for reading sub-skills and reading comprehension. All readers were trained to identify the voices of four English talkers and four French talkers during training phases. Following training, all readers were tested on retention of the trained sentences as well as generalization to novel sentences produced by the same talkers. The results indicated that compared to the average readers, the advanced readers (1) showed higher talker identification during training for both the native and non-native talkers, (2) required less exposure to learn the non-native voices, and (3) generalized the non-native voices to a greater degree. These results extend findings from previous research to include a gradient effect of language competence on talker identification even among within-normal differences in reading ability.

Major Advisor

Rachel M. Theodore

Share

COinS