Abstract
Three experiments examined acoustic and perceptual characteristics of the speech of gay; Lesbian; Bisexual (GLB) and heterosexual people. Experiment 1 examined the acoustic characteristics of single words produced by both men and women who identified as either GLB or heterosexual. The largest differences between GLB and heterosexual women were in the F1 frequency of /ε/ and the F2 frequency of /oυ{phonetic}/. The largest differences between the groups of men were in the F1 frequency of /ε/ and /æ/, and the spectral skewness of the fricative /s/. Experiment 2 showed that listeners' judgments of perceived sexual orientation were related to the acoustic parameters found to differ in Experiment 1: Listeners showed greater sensitivity to differences in men's sexual orientation when listening to words containing low front vowels and than when listening to words containing back vowels. Moreover, Regression analyses showed that judgments of men's sexual orientation were influenced by /s/ skewness, the F1 frequency of low front vowels, and the F2 frequency of back vowels. Judgments of women's sexual orientation were predicted most strongly by the F1 frequency of low front vowels and the F2 frequency of back vowels. Experiment 3 showed that the judgments of perceived sexual orientation collected in Experiment 2 were strongly related to judgments of perceived height and perceived speech clarity made by independent groups of listeners. Taken together, the results provide a more comprehensive picture of the acoustic and perceptual characteristics of GLB speech styles than has been provided by previous research. Moreover, the results of Experiment 3 suggest that listeners' percepts of GLB speech styles may be related to their perception of other speech characteristics.
Original language | English (US) |
---|---|
Pages (from-to) | 202-240 |
Number of pages | 39 |
Journal | Journal of Phonetics |
Volume | 34 |
Issue number | 2 |
DOIs | |
State | Published - Apr 2006 |
Bibliographical note
Funding Information:This research was supported by a Grant-in-Aid of Research, Scholarship, and Artistry from the University of Minnesota Graduate School. The authors are extremely grateful to Sarah Anderson, Molly Babel, Frank Davis, and Micaela Zolltheis for their help in analyzing these data. We gratefully acknowledge Bert Remijsen for describing the procedure for calculating H1−H2, and for providing Praat scripts to implement it. We also thank Pauline Welby for help with the other Praat scripts used in this investigation, and Edward Carney both for help in calculating the nonlinear regressions described in Section 2.4.3 , and for fitting the Bradley–Terry models described in Section 4.4.1.2 . This research has been presented previously at the Fall 2003 biannual meeting of the Acoustical Society of America in Austin, TX, the Fall 2004 meeting of the American Speech–Language–Hearing Association, and in a number of invited colloquia. We gratefully acknowledge the comments provided by audiences at those locations. We are particularly grateful to comments and feedback from Molly Babel, Mary Beckman, Ann Bradlow, Sarah Jefferson, Henrietta Jonas-Cedergren, Miriam Krause, Norma Mendoza-Denton, Janet Pierrehumbert, and Pauline Welby.