SASS Lab Projects

Ongoing Projects

Project FYI: Finding Your Identity in speech and language

Phonetic differences between adult men and women are pervasive, and are generally claimed to result from sexual dimorphism in the larynx and vocal tract (Fitch & Giedd, 1999; Titze, 1989). This dimorphism does not appear reliably until puberty (Barbier et al.., 2015; Vorperian et al., 2011) It is surprising, then, that naïve listeners provide different ratings for the content-neutral speech of prepubertal children assigned male at birth (AMAB) and assigned female at birth (AFAB) on scales that range from “definitely a male” to “definitely a female” (Barreda & Assmann, 2021; Fung et al., 2021; Munson, Koeppe, & Lackas, 2022; Perry et al., 2001). Acoustic analyses of children’s productions suggest that gendered speech is achieved not through global emulation of adult male and female ways of speaking, but by selective learning of salient gendered phonetic variants in the ambient language (Foulkes et al., 2005; Munson et al., 2015; Wong, Koeppe, Cychosz, & Munson, 2022). The nature and extent of this selective learning is a model for understanding how children’s emerging sense of self focuses their language acquisition, both in naturalist learning, and in clinical and educational settings. In this project, we are conducting a series of experiments, and analyses of naturalistic data, to understand how, why, and when children with and without speech and language disorders learn gendered ways of speaking.

Project REI: Race and Ethnicity in Intelligibility

Because of changing US demographics, communication between older individuals with hearing impairment and younger individuals—including those in caregiving and service-delivery roles—is increasingly likely to be between individuals of different races and ethnicities. Research has shown that a talker's speech can become less intelligible when they become aware of a talker's race or ethnicity by seeing a picture of the talker. This project builds a new corpus of audiovisual speech stimuli produced by individuals from diverse races and ethnicities. We are using these in a series of intelligibility experiments with younger listeners with normal hearing, older listeners with normal hearing, and older listeners with hearing impairment. We are also collect information on individuals' attitudes toward individuals of different races and ethnicities, as well as the ethnic and racial diversity of their peer groups. Results from this project will help us better understand the mechanisms that underlie effects of race and ethnicity on speech intelligibility, and whether the speech perception of older adults with hearing impairment is disproportionately poorer in interactions with racially and ethnically diverse interlocutors. Funding: NIH grant R21 DC018070 to Munson

Project SECSR: Social Effects on Children's Speech Recognition

Adult listeners associate linguistic variation, like variation in word choice and pronunciation with social features like age, race, and gender, among others. This influence is multi-faceted: language comprehenders use language variation to categorize language producers socially. Once an comprehender has categorized a producer a particular way, they apply a set of expectations about how the producer will say things they haven't yet said. For example, once a comprehender has classified a producer as a cisgender middle-class white woman, they might expect that individual to produce the "s" sound a particular way, in line with their stereotypes about cisgender However, comparatively little is known about how this kind of sociolinguistic knowledge develops in young children. If social beliefs impact language comprehension the way that they do in adults, then we would predict that social effects on expectations would affect children's language comprehension, both in everyday interaction with peers and caregivers, and in high-stakes interactions with individuals who hold power over them. This research examines how the influence of social expectations on language comprehension affect children's performance in one of the highest-stakes interactions they have, standardized educational tests. Gaps in vocabulary test scores between Black and white children entering school have previously been interpreted as evidence of impoverished Black linguistic experiences. However, the failure of these linguistic assessments to explicitly test sociolinguistic expectations may obscure the strengths of children learning non-mainstream language varieties. The planned experiment presents 4-6 year old children with a modified version of one widely used assessment, Peabody Picture Vocabulary Test, which is part of almost every assessment of children's language. This study measures whether children's perception of themselves and of the test administrators significantly influences test scores. That is, we examine whether children perform better when they interact with preferred test examiners. If we can demonstrate that differences in vocabulary scores can be partly explained by children predictably exhibiting different sensitivities to the test administrator’s appearance and linguistic behavior, then we will have provided a new explanation for part of the ‘readiness gap’ that does not rely on assuming that it stems solely from deficits in Black children’s knowledge.

Project POV: A Palette of Voices for transmasculine individuals

There is not a strong research basis for gender-affirming speech and voice services, especially for transmasculine individuals. The project in this application is part of rethinking and enriching gender-affirming speech and voice services. In it, we are building and validating a clinical tool that SLPs can use in the context of client-centered, culturally responsive speech habilitation for trans individuals. The specific activity that we propose is the development of a palette of voices (POV) for transmasculine individuals. The POV will have speech samples that are collected from a racially and culturally diverse population, and which include both natural productions, and productions that have been acoustically modified to represent the many different ways of sounding like a man (or, perhaps, sounding like a gender that is neither exclusively male nor exclusively female). This project involves building, validating, and field-testing this tool.Funded by a seed grant from the University of Minnesota College of Liberal Arts

Recently Completed Projects

SKiLLS4Words

Small Kids Learning Language Sounds for Words. This project was a longitudinal study of phonological development and vocabulary growth. The data are interesting in their own right, and provide a fertile ground for studying lots of interesting things about speech sound development. Funding: NIH grant R01 DC02932 to Jan Edwards, Mary E. Beckman, and Ben Munson

Math to Mouth

This project examined how children learn the relationship between the way they talk and the way they sound to others. In this project, we collected lots of data on how people perceive children's speech. This has led us to develop new tools noting the accuracy of children's speech that go beyond phonetic transcription.Funding: NSF Grants BCS-0729306 [Mary E. Beckman & Eric Fossier-Lussier], BCS-0729140 [Jan Edwards], and BCS-0729277 [Ben Munson]

Page updated

Report abuse