Creaky, She Spoke: Examining f0, Vocal Creak, and Perceptions of Young Women’s Professionalism
Recent sociolinguistic work on creak (also known as vocal fry, vocal creak, and creaky voice) has generally focused on its gendered use and perceptions, and more specifically on creak and young women. While Yuasa (2010) found creaky young women to be perceived as “educated, professional, and upwardly mobile” (p.316) sounding, Anderson et al. (2014) and Gallino and Pinto (2021) found the opposite in that creaky young women were perceived to be “less competent, less educated, less trustworthy, less attractive, and less hireable” (p.5). However, all of these works consider creak in isolation, eschewing analyses that examine creak in tandem with other linguistic variables. As voices are perceived as a whole unit, and not singular linguistic features, this leads to an incomplete understanding of the ways in which listeners might perceive creak differently when it is paired with other features such as f0, prosody, and phonetic variation related to dialect, etc. Work examining creak interacting with other linguistic features will provide better understanding for how, when, and why creak is perceived in specific ways by listeners. This study interrogates whether there is an interaction between f0 and creak that might affect professionalism perceptions of speakers and, additionally, how a listeners’ speech attitudes about women and creak might mediate f0 and creak’s effects on professionalism ratings. 125 participants rated stimuli produced by five young white women from the Midland region sourced from the Nationwide Speech Project corpus (Clopper and Pisoni 2004) on six Likert scales: professional, attractive, friendly, feminine, educated, and authoritative. Additionally, participants completed a set of workplace and women’s speech attitudes Likert scale measures (e.g., “to be successful in the workplace, young women should change how they speak to sound more professional”). The presence of creak in stimuli was determined via impressionistic listening and examining spectrograms in Praat. Factor analysis showed that ratings for the six traits patterned into two factors: “Competence” and “Warmth”. These two factors were used as dependent measures in linear mixed effects regression models, with average ratings for both factors as the dependent variables. The Competence model included main effects for Workplace Sexism attitudes (p<.001) and f0 (p<.01), but failed to reproduce findings from the aforementioned previous studies with regards to creak. This study supports previous work by Parker and Borie (2017) that calls for more nuanced and complex analyses of creak moving forward, and the need to move beyond attempting to examine creak as a variable in isolation. It also points to the possibility that the indeterminacy of previous findings on creak and social meaning can be attributed, in part, to differences in experimental design.