|
|
 E-mail this message to a friend
|
|
Title:
|
Individual Differences and the Link between Speech Perception and Speech Production
|
|
Author:
|
Rochelle Newman
|
|
Email:
|
click here to access email
|
|
Homepage:
|
http://www.bsos.umd.edu/hesp/newmanr.html
|
|
Degree Awarded:
|
University at Buffalo
, Department of Psychology
|
|
Degree Date:
|
1997
|
|
Linguistic Subfield(s):
|
Psycholinguistics
|
|
Director(s):
|
James Sawusch
|
|
|
Abstract:
|
|
A long-standing question in speech research concerns the degree of interrelation between speech perception and speech production. That is, are the representations used for these two different processes tightly linked, or possibly even identical? A related issue is whether there are reliable differences between individuals' perception which are related to the idiosyncrasies of their production.
Motor Theory (Liberman, Cooper, Harris & MacNeilage, 1962; Liberman & Mattingly, 1985) first proposed that speech perception takes place in reference to production. This would mean that the perceptual process makes use of the representations developed for production, and that differences between individual's productions should be reflected in their perception, as well.
A number of experiments have attempted to examine this issue over the years, but results have been quite variable. It is unclear whether this confusion is because the effect itself is variable, or whether more sophisticated experimental techniques might resolve the issue. The present set of experiments was designed to investigate this topic more closely.
The experiments reported here are modeled after an experiment by Miller and Volaitis (1989) in which they asked subjects to rate members of a series for their category goodness. This allowed them to examine perceptual 'prototypes' of a phoneme category for an individual listener. In the experiments described here, these perceptual prototypes were correlated with acoustic measurements of each listener's own productions . In the first experiment, listeners were asked to rate members of a VOT series ranging from /ba/ to /pa/ to /*pa/ (beyond a good 'p')Individuals who preferred tokens of /p/ with longer VOTs also produced longer VOTs in their own productions. Additional variance in the perceptual prototype was explained by production of /ba/. This suggests that voiced and voiceless stops provide separate, non-overlapping information about individual's mental representations, and that differences in perception are related to differences in production. A final finding from this experiment was that individual's perceptual prototypes tended to have more extreme VOT values than their own productions. That is, individuals seemed to demonstrate a 'hyperarticulation' effect, as has been previously shown for vowels (Frieda, 1997; Johnson et al., 1993).
In Experiment 2, neither centroid of frication nor formant frequencies at onset of vocal pulsing demonstrated any correlation between perception and production in a /s/-/*/ series. In the third experiment, a number of proposed cues were examined for stop consonants differing in place of articulation. Locus equations demonstrated no correlation between the two modalities for /b/, /d/, and /g/. Spectral moments and spectral peak differences showed no significant correlations on individual submeasures, but canonical correlations examining these entire sets of cues yielded high correlations. These canonical correlations were equal in size for the two sets of cues, suggesting that the sets are approximately equivalent descriptions of the information that listeners actually use.
The results from the set of experiments are not as clear as might be desired. The significant effect in Experiment 1 suggests that some links between perception and production do exist, and can be found with a suitable methodology. However, the variability across experiments suggests that this link is not especially strong, arguing against the notion that the modalities might share the same representations. Rather, it appears more likely that the link is indirect. Since the voice individuals have the most experience hearing tends to be their own, individuals' productions are likely to have a substantial influence on their perceptual prototypes.
|
|
|
|
|
|
Page Updated: 24-Nov-2009
Please report any bad links or misclassified data
LINGUIST Homepage | Read
LINGUIST | Contact us
While the LINGUIST List makes every effort to ensure the linguistic relevance of sites listed
on its pages, it cannot vouch for their contents.
|
|
