You are here

FAU Collections » FAU Research Repository » FAU Theses and Dissertations

Perceptuomotor biases in vowel imitation

Download pdf | Full Screen View

Date Issued:
2003
Summary:
The method of imitation provides an alternative approach to longstanding problems in speech research such as the organization of the vowel space and its relation to linguistic categories. Imitation of self-produced targets is particularly useful because (a) it allows a subject to report the quality of a multidimensional percept quickly and efficiently and (b) the subject is in principle capable of matching the target perfectly, so systematic deviations in the imitation can be ascribed to perceptuomotor biases. The self-imitation task was used to study perceptuomotor bias in the F1 x F2 vowel spaces of native speakers of American English, using two kinds of tasks. In "Serial Imitation", the imitations of a subject were played back to him for further imitation, thereby creating a set of trajectories in the vowel space. In "Multiple Imitation", each target was imitated several times, thereby allowing the perceptuomotor bias and variability to be calculated. Both tasks showed that subjects were consistently inaccurate in imitating themselves and that different subjects had different patterns of bias directions over the vowel space. The bias did not seem to be influenced by the temporal structure of the target nor by the subject's linguistic background and phonetic training. Articulatory modeling further showed that the bias was not due to low-level noise during production. The cause of the bias is posited to be the subject's psychological set, viz. his or her tacit notion of what aspects of the target are to be attended to and emphasized. An additional result was that while the bias directions differed across subjects, the distribution of bias magnitude and formant variability were remarkably consistent across subjects and sessions. In particular, F1 and F2 varied independently of each other and the formant standard deviations were similar to the formant discrimination limen. It is proposed that the psychological set is akin to a cognitive "policy" that can vary across subjects and sessions, while the underlying mechanism that carries out the policy is relatively unchanging.
Title: Perceptuomotor biases in vowel imitation.
113 views
30 downloads
Name(s): Vallabha, Gautam Kumar
Florida Atlantic University, Degree Grantor
Charles E. Schmidt College of Science
Center for Complex Systems and Brain Sciences
Type of Resource: text
Genre: Electronic Thesis Or Dissertation
Issuance: monographic
Date Issued: 2003
Publisher: Florida Atlantic University
Place of Publication: Boca Raton, Fla.
Physical Form: application/pdf
Extent: 202 p.
Language(s): English
Summary: The method of imitation provides an alternative approach to longstanding problems in speech research such as the organization of the vowel space and its relation to linguistic categories. Imitation of self-produced targets is particularly useful because (a) it allows a subject to report the quality of a multidimensional percept quickly and efficiently and (b) the subject is in principle capable of matching the target perfectly, so systematic deviations in the imitation can be ascribed to perceptuomotor biases. The self-imitation task was used to study perceptuomotor bias in the F1 x F2 vowel spaces of native speakers of American English, using two kinds of tasks. In "Serial Imitation", the imitations of a subject were played back to him for further imitation, thereby creating a set of trajectories in the vowel space. In "Multiple Imitation", each target was imitated several times, thereby allowing the perceptuomotor bias and variability to be calculated. Both tasks showed that subjects were consistently inaccurate in imitating themselves and that different subjects had different patterns of bias directions over the vowel space. The bias did not seem to be influenced by the temporal structure of the target nor by the subject's linguistic background and phonetic training. Articulatory modeling further showed that the bias was not due to low-level noise during production. The cause of the bias is posited to be the subject's psychological set, viz. his or her tacit notion of what aspects of the target are to be attended to and emphasized. An additional result was that while the bias directions differed across subjects, the distribution of bias magnitude and formant variability were remarkably consistent across subjects and sessions. In particular, F1 and F2 varied independently of each other and the formant standard deviations were similar to the formant discrimination limen. It is proposed that the psychological set is akin to a cognitive "policy" that can vary across subjects and sessions, while the underlying mechanism that carries out the policy is relatively unchanging.
Identifier: 9780496283651 (isbn), 12028 (digitool), FADT12028 (IID), fau:8943 (fedora)
Collection: FAU Electronic Theses and Dissertations Collection
Note(s): Adviser: Betty Tuller.
Thesis (Ph.D.)--Florida Atlantic University, 2003.
Subject(s): Psychology, Cognitive
Held by: Florida Atlantic University Libraries
Persistent Link to This Record: http://purl.flvc.org/fcla/dt/12028
Sublocation: Digital Library
Use and Reproduction: Copyright © is held by the author, with permission granted to Florida Atlantic University to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Use and Reproduction: http://rightsstatements.org/vocab/InC/1.0/
Host Institution: FAU
Is Part of Series: Florida Atlantic University Digital Library Collections.