Colloquium speaker: Eric Russell Webb
‘Your /y/ is so gay, but your /s/ is so straight’: Exploring the phonetic encoding of orientation
Sociophonetic research suggests that speech encodes linguistic and extra-linguistic information, including gender, group identity, socio-cultural provenance, and religiosity. Beginning in the 1980s, attention has been given to sexual orientation, notably to what distinguishes GLB from straight/heterosexual speakers. This work has, however, focused nearly exclusively on English, notably on North American varieties, and has also demonstrated a propensity to determinism, specifically the assumption of an inevitable connection between phonetic style and psychological identity.
In this talk, I explore the phonetics of orientation in three parts. Firstly, I review existing literature on gay speech. Noting the significance of previous work, I briefly discuss several shortcomings, especially the methodological predicates and assumptions. I next overview a multi-stage project investigating the phonetic encoding of identity in French, Italian, and Dutch, as well as the indexicality and salience of these within and across languages and. Most of the talk is given to a close analysis of French data and to the phonetic features associated with shifting from neutral to straight and gay speech styles.
This study represents a departure from previous work, notably in methodology. Rather than beginning with the expectation that gay speakers employ a gay style and straight speakers a straight one, I purposefully ignore talker self-identity. Participants included in the study were screened and accepted solely based on self-reported comfort with gay- and straight-identified individuals, as well as frequency of interaction with members of both subgroups. Talkers were asked to read aloud three paragraphs employing distinct styles (introductory, scientific, and narrative retell), all of which had been blanched for obvious lexical associations with one or another orientation. After a neutral reading, talkers were asked to read these in such a way as to assure that listeners would perceive them as straight and gay, respectively. The study thus aims to understand the phonetic indices of both gay and straight speech styles, within and across speakers, as these contrast to neutral style.
For the experiment reviewed in this talk, results yielded approximately 4.5 minutes of recorded speech, containing approximately 7000 measured data points, per individual. Acoustic analysis using Praat focused on 13 characteristics: sibiliant duration and quality (peak and peak downstep), lateral duration and quality, oral vowel duration and quality, nasal vowel duration, pitch/F0 envelop, intonation contours on focal elements, intensity window, duration of focal elements, schwa elision, and liaison. Of these, inter-speaker data indicate that only a subset of features are used uniformly to convey gay style: nasal vowel duration, the duration of high rounded vowels /y, u/, pitch envelop, and intonation contours. Straight style appears to be closely associated with vowel quality, decreased /s/ duration, and divergent focus contours. Additionally, intra-speaker evidence suggests that individuals make differential use of a menu of phonetic possibilities (viz. Zwicky 1997), notably vowel quality and focus duration, for both gay and straight styles, and that the degree to which these are employed varies according to speech task (scientific readings being the least and introductory readings the most divergent). Importantly, French appears to differ significantly from English in that /s/ quality and /l/ duration and quality do not significantly diverge across styles or tasks.