This post was written by Brandon Merritt, a Ph.D. candidate in the Department of Speech, Language, and Hearing Sciences at Indiana University.
When we hear a voice, the speaker’s gender is one of the first things we notice about them. Even if we’re asked to categorize speakers based on dialect or accent, we still gravitate toward grouping them based on how we perceive their gender.
In my lab, we study which specific physical properties of speech (known as acoustic cues) control how gender identity is perceived. The acoustic cues we examine are voice pitch (controlled by how quickly the vocal folds vibrate), voice resonance (a result of the size and shape of the throat and mouth), articulation (how we produce the consonant and vowel sounds in our speech), and intonation (the melody we produce with our speech).
In read speech or in conversation, articulation and intonation appear to become more important indicators of gender identity. Articulation, or how we produce consonants such as “s” and “sh” and vowel sounds, seems to be an especially strong cue to gender for listeners. Speakers who make more precise consonants and vowels may be perceived as more feminine, while those who use less precise articulation may be perceived as more masculine.
Unfortunately, research on how gender is projected and perceived through speech has historically been limited to cisgender individuals (i.e., those whose gender identity matches the sex assigned to them at birth), and excludes transgender individuals (i.e., those whose gender identity does not match the sex assigned to them at birth). So, most of what we know about gender and speech is based off of cisgender men and women and assumes a binary classification of gender (i.e., male or female).
Because the number of people who identify as transgender across the world is growing, listeners are likely exposed to speakers with a wider range of gender identities than only cisgender male and cisgender female. Therefore, excluding transgender speakers from this research neglects the increasing gender diversity of contemporary society. Furthermore, studying this topic can help guide people if they want to intentionally alter how they speak to better represent their gender identity.
Recent work in our lab is helping to fill this gap. We are exploring what aspects of a person’s speech are most important to listeners when making judgements about gender. We collect speech samples from cisgender and transgender speakers and systematically manipulate what acoustic cues listeners hear from these speakers. We can then examine how a listener categorizes both a speaker’s gender identity and masculinity/femininity and compare which acoustic cues are most important for a listener’s judgement.
These data allow us to examine how speakers combine different sorts of speech features to express gender identity, including what we may think of as typical male or female identities, as well as more gender-diverse identities, such as non-binary or genderqueer. Speakers could, for instance, selectively choose to speak with a lower voice pitch (a typically masculine characteristic), but include more clearly produced consonant sounds and make greater use of musicality of voice (typically feminine characteristics) to create speech that counters notions of typical “male” or “female” speech and allows for more nuanced gender expression.
Our future work will examine the specific articulatory cues that may be especially important for gender judgements (e.g., what sorts of consonant and vowel productions provide the strongest cues to gender?). Also, we will explore how listeners’ experiences may shape their perceptions of a speaker’s gender (e.g., do listeners with social networks more inclusive of gender-diverse people have different criteria for categorizing a speaker’s gender?).
A clearer understanding of the projection and perception of gender in speech provides guidance for individuals seeking to better align their voice and speech with their gender identity.
Edited by Taylor Woodward and Evan Leake