Poster #197 - Who Said That? Accents Guide Children’s Inferences About the Speaker’s Race

• In Event: 3-003 - Poster Session 09
  In Poster Session: PS 09 Section - Social Cognition

Integrative Statement

The speech signal not only conveys a linguistic message, but also discloses aspects of the speaker’s physical (e.g., sex, age), psychological (e.g., mood), and social (e.g., regional origin) background. Successful comprehension of spoken language involves the integration of all types of information (Borovsky & Creel, 2014), thus raising the question of how children learn to process speakers’ voices. Research in this domain has revealed an early sensitivity to speaker-specific cues (DeCasper & Fifer, 1980), but it is unclear what assumptions young children make about the speaker based on voice information alone.

Here, we examine children’s mapping of unfamiliar voices to faces. In particular, we ask whether children use speakers’ accents to infer their race. Although certainly not infallible, adults conceivably make predictions about a speaker’s race from accents alone (Purnell et al., 1999). Children, too, have been shown to be sensitive to accents – in particular to infer regional background (Weatherhead et al., 2016; 2017) and make social judgments (Kinzler et al., 2007). But do they track the probabilistic relationships between accent and race?

To address this question, Caucasian monolingual English-learning children (N=36) were tested in Experiment 1. They were between 4;0 and 7;6 years of age, and were reported to have little exposure to English produced by non-US English speakers. Children were tested on a two-alternative forced choice task. On each trial, they saw one East-Asian and one Caucasian face (matched for sex, age, attractiveness, and happiness; Ma et al., 2015) side-by-side on a screen (see Figure 1). They then heard an English sentence produced by (i) a native English speaker of the regionally dominant accent, (ii) a Mandarin-accented English speaker, or (iii) an Indian-accented English speaker. Children were asked to point to the face that best represented the speaker. Results revealed that the proportion of East-Asian face selections depended on the accent of the speaker (p < .01), with children being more likely to choose the East-Asian face when the speaker had either a Chinese or an Indian accent (ps < .01, see Figure 2a).

Given that both unfamiliar accents triggered a large amount of East-Asian face selections, Experiment 2 was conducted to determine whether children associate particular accents with particular races or whether they consider any non-local English accent as unfamiliar and link these unfamiliar accents to other-race individuals. To this end, Experiment 1 was replicated (N=36, identical selection criteria), replacing the Indian accent with a Scottish accent. In line with the findings from Experiment 1, the selection of East-Asian faces depended on accent (p < .01). Crucially, however, unlike the Indian accent, the Scottish accent generated fewer East-Asian face selections than the native accent (ps < .01, see Figure 2b).

Taken together, this work shows that despite it being a non-deterministic cue, children use a speaker’s accent to draw race-based inferences. Moreover, different accents lead to different inferences, suggesting that this association is at least somewhat accent-specific. Early on in life, accents thus start functioning as a social grouping variable that guides children’s expectations about speaker identity.

Authors

• Genevieve Franck, University at Buffalo
  Presenting Author

• Michelle Katharine Tulloch, Florida Atlantic University
  Non-Presenting Author

• LiPing Lin, University at Buffalo
  Non-Presenting Author

• Marieke van Heugten, University at Buffalo, The State University of New York
  Non-Presenting Author