Inter-Talker Variability and Perception of Phonetically Similar Tones
L2 speech perception
lexical tone perception
lexical tone production
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
EmbargoRelease after 03-May-2019
AbstractThis dissertation probes the processing of phonetically similar tones under high-variability conditions (i.e., with inter-talker variability), as well as whether orthographic cues for tones affect the perception of non-native tonal contrasts. This current work consists of three studies. Throughout these studies, five Cantonese tone pairs were examined: 1) Mid-Level vs. Low-Level, 2) High-Rising vs. Mid-Rising, 3) Low-Falling vs. Low-Rising, 4) Low-Falling vs. Low-Level, and 5) Low-Rising vs. Low-Level. The phonetic contrast in the first pair is f0 height. The contrast in the second pair is the magnitude of f0 rise. The contrast in the last three pairs is overall f0 contour. Chapter 2 reports on native perception and production of the above-mentioned tones. Through a high-variability mixed-talker AXB task and a mixed-talker delayed shadowing task, I found that the f0-height difference in the two rising tones was easier to perceive than that in the two level tones. In addition, f0-contour differences were easier to perceive than f0-height differences. I also found that participants accommodated the model talkers when shadowing words from them: participants adjusted the average f0 height for the level tones and the magnitude of f0 rise for the rising tones according to the talkers. Chapter 3 reports on non-native perception of lexical tones at first exposure. Three groups of non-Cantonese speakers (English, Mandarin, and Bangkok Thai) completed a mixed-talker AXB task that tested discrimination performance and a mixed-talker delayed shadowing task that tested identification and production. Results showed that, for all non-Cantonese speakers, the perception of the f0-height difference in the two rising tones and that in the two level tones were equally difficult, and that f0-contour differences were easier to perceive than f0-height differences. In addition, L1-dependent patterns were found. In general, tonal-language speakers perceived f0-contour differences better than English speakers. English and Mandarin speakers outperformed Thai speakers in perceiving two rising tones that have similar contour shapes but different f0 heights. Unlike native listeners, non-native listeners were less successful in using syllable-extrinsic information for perceiving f0-height contrasts. Chapter 4 reports on the effect of orthographic cues for tones on non-native perception of tonal contrasts. The non-Cantonese speakers (reported in Chapter 3) who were in an Auditory-Visual training group saw iconic tone letters as immediate feedback in a mixed-talker AX training task, whereas those in the Auditory-Only group did not. Results from the training tasks revealed that the training with tone letters was more beneficial, as there was a facilitation effect on a larger scale, affecting more learners and the learning of more tonal contrasts: the tone letters facilitated the learning of the f0-height contrast between two level tones for all three groups of learners, as well as the learning of f0-contour contrasts for English speakers. In sum, this dissertation provides evidence that 1) the perception of tones contrasting in various phonetic dimensions is affected differently by inter-talker variability; 2) the perception of non-native tones is affected by native language background; and 3) the effect of iconic orthography varies as function of the L1 background of the learners and the specific L2 contrasts tested.
Degree ProgramGraduate College