Valerie Fridland                         Kathryn Bartlett

University of Nevada, Reno                    University of Nevada, Reno

 

 

Audio | Handout | Figures

What We Hear and What it Expresses: The Perception and Meaning of Vowel Differences Among Dialects

The effect of vowel quality on listeners’ perceptions has received little attention though much has been made of how changes in vowel quality function as a means of symbolic identity, uniting and dividing groups of speakers. While the patterned use of linguistic variants by different groups within communities appears to suggest a paralinguistic social function, actual empirical research measuring the role of perception in assigning meaning to variation is scarce. In addition, how speakers’ own productive system affects their accuracy in perceiving and evaluating vowel differences is an area of study rarely examined, although such exploration could contribute much to our theories of speech production and perception.

Southern speech is a highly stereotyped and very salient dialect that is undergoing massive changes to the vowel system, changes that are in some cases regionally unique (as with front vowel changes), but with some, such as back vowel fronting, shared more generally across regions. Such local and national contrasts in aspects of vowel production serve as a good starting point in the examination of how distinguishing vowel changes are perceived and evaluated compared to those changes that are not inter- and intra-regionally defining. To this end, the current paper is designed to study speakers’ perceptual awareness and social evaluation of specific regional vowel variants using acoustically manipulated speech samples. While experimental in design, this study provides a unique and innovative method of measuring speakers’ sensitivity to slight changes in formant position and how such subtle phonetic changes are indeed used as socially salient categorization cues by speakers.

The results are based on 175 African-American and European-American respondents from Tennessee and 150 European-American respondents from Nevada, Utah and California who listened to a series of ‘different’ male and female guises. Each guise was a monosyllabic token with the vowel synthesized to approximate shifted positions in Southern and Northern regional shifts. Only F1 and F2 formant structure differed among guises. In a three part test, participants rated each ‘speaker’ on a semantic differential scale for four factors; degree of Southernness, degree of difference between shifted tokens of the same vowel, and levels of education and of pleasantness. Anovas and paired comparisons t-tests will be run to determine accuracy in the selection of the most “Southern” guises, whether accuracy is affected by the degree of participants’ productive participation in the various aspects of the shifts, whether the results show distinctions by participants’ region, ethnicity, gender or age, and whether different education and pleasantness scores are assigned on the basis of vowel position. The data from the Western subjects will be compared to the Southern subjects to determine how exposure to specific formant ranges of vowel variants affects respondents’ ratings and accuracy.

The research aims to contribute both to the larger question of how much fluidity speakers’ have in adjusting their perceptions based on exposure to different vowel frequency ranges and how very low-level phonetic information is used by speakers and, more narrowly, to the question of how extensively local Southern norms affect the adoption of incoming changes.