The acoustics of lesbian vowels and the perception of womenís sexual orientation

Tessa Bent, Janet Pierrehumbert and Michael Bailey

The identification of a personís sexual orientation based on their speech alone seems to be a robust phenomenon, yet what characterizes the speech of gay and straight speech, specifically the acoustics, is an area which little is known. A few studies (Terango 1966, Lerman and Damste 1969, Gaudio 1994, Avery and Liss 1996) have investigated the acoustics of gay male speech (also less masculine sounding male speech and effeminate speech) but almost nothing is known about the acoustics of lesbian speech (except see Moonwomon 1985, 1997).

In a previous study Bailey and colleagues demonstrated that naïve listeners were able to identify the sexual orientation of women solely by listening to recordings of their voices. They played 80 listeners recordings of four sentences read by 17 heterosexual women and 33 lesbian and bisexual women. The listeners rated the "gayness" of the voice on a scale of 1 to 7 where "1" is "sounds totally straight" and "7" is "sounds totally gay/lesbian." The perception data showed that listeners were able to identify the sexual orientation of the women based exclusively on the recordings of their voices. In the current study, we are investigating the following questions: (1) what acoustic cues are listeners using to make the identification of sexual orientation? (2) what is the relationship between the cues listeners attend to in making judgements of sexual orientation and the extent to which these cues are realized in lesbian speech? We are measuring whether straight women and lesbians produce their vowels differently and if vowel formant frequencies are related to how "gay" sounding listeners rated the womenís voices. The first, second and third formants were measured for the vowels /i/, /e/ /ae/, /u/ and /a/. Preliminary results show that generally homosexual women have lower first and second formant values than heterosexual women. Additionally, perceived gayness and at least one of the formant frequencies were significantly correlated for all of the vowels. Therefore, we conclude that homosexual women are attempting to achieve the sound of a longer vocal tract as evidenced by the lower first and second formants.