Phonetic Cues and the Perception of Gender and Sexual Orientation

Phonetic variation is a well-established type of cue used in the construction and perception of social identities such as gender, concepts of masculinity and femininity, and sexual orientation. Less established is how cues indexed with inter-related identities are integrated or weighted relative to each other. Because the cues identifying gender and sexual orientation are potentially overlapping, (i.e., gay and lesbian speakers are often perceived as using cues indexed with the opposite gender), in this study, experimental methods are used to assess cue effects. Experimental cues used are fundamental frequency (f0), first and second vowel formants (Fn), and the center of gravity of the fricative /s/. These were chosen as all were significant in the perception of sex, gender and/or sexual orientation, or in terms of production by such groups in previous studies.;In this study's experiments, three groups of participants (n=105) heard target stimuli consisting of /alpha/, /salpha/ or /alphas/, created by combining altered and unaltered cues. Each participant heard a stimulus and chose a perceived gender and sexual orientation for that stimulus. Results indicate that the outcomes for the dependent variables of Perceived Gender, Men's Perceived Sexual Orientation (MPSO) and Women's Perceived Sexual Orientation (WPSO) result from the same cues but that cue integration strategies vary. The area of greatest overlap is for Perceived Gender and MPSO with F0 Up cues resulting in more women and gay men percepts, and FN Down cues resulting in more men and straight men percepts. For WPSO results the vocalic cues strongly affect each other, although F0 Up cues tend to result in more straight women percepts. These results suggest that listeners use experienced and stereotypic cues when determining gender and sexual orientation. The results also suggest that listeners use an androcentric template for all three dependent variables. Cue variants are not marked with specific indexical meanings such as "women" or "femininity", but are marked in terms of man-based norms. Crucially, cue integration involves both social markedness based on stereotypes and phonetic markedness based on cue ranges. Cue Integration Theory is proposed to account for the empirical results for all three dependent variables.