A study by researchers in the Linguistics Program and the Department of Psychological & Brain Sciences finds that the race of a speaker doesn’t affect comprehension — challenging a prominent study on the topic.
When we have a face-to-face conversation with someone, we’re taking in more than just the sound of their voice. “It’s generally accepted that the way a person looks can affect the way you process their speech,” said Kristin Van Engen, an associate professor in the Linguistics Program and Department of Psychological & Brain Sciences.
But a new study published in Language and Speech offered a surprising twist: Studying several accents and racial types, researchers found that a speaker’s racial features had minimal impact on a listener’s ability to comprehend their voice. This is good news, Van Engen said. “We should be glad that people didn’t have trouble understanding language just because of the race of the speaker.”
The study was co-authored by Van Engen and Drew McLaughlin, a former WashU PhD student who is now a postdoctoral researcher at the Basque Center on Cognition, Brain and Language in San Sabastián, Spain.
Earlier research by other scientists had suggested that listeners are particularly sensitive to — and occasionally flummoxed by — the race of a speaker. For example, an oft-cited study from 1992 found that American students had more trouble understanding a lecture in American-accented English if they were shown a face with East Asian features. “People have taken these original findings and run with them as a sign of racial bias,” Van Engen said.
Van Engen noted that another previous study found that English-speaking American listeners better understood Mandarin-accented speech when shown an East Asian face. That result, she said, suggested that racial information could shape listeners’ expectations of what a given speaker will sound like.
Van Engen and McLaughlin recruited 640 white adults for the study. The subjects, who were all native English speakers, were asked to transcribe snippets of English-language speech while looking at a picture of the presumed speaker. The audio clips were presented in two different accents: Arabic and Mandarin. The images included a white woman, a woman with East Asian features, and a woman with Middle Eastern features wearing a hijab. A fourth image was a silhouette with no racial features.
Contrary to expectations, the racial features of the speaker had almost no impact — positive or negative — on the listeners’ comprehension. Early results seemed to suggest that subjects had less trouble understanding Mandarin-accented English when the speaker appeared to be East Asian, but that effect disappeared as the study went on.
The study did show some subtle trends that may warrant more investigation. Subjects had slightly more trouble understanding Arabic-accented speech if the speaker seemed to be East Asian but, surprisingly, they seemed to understand a Mandarin accent a bit better if the speaker appeared to be Middle Eastern. “The results are complex and nuanced,” Van Engen said.
Van Engen suspects that many factors may affect a listener’s ability to comprehend speech in an unfamiliar accent. People who have a lot of experience with a particular accent or who have been exposed to many different accents may be less likely to be thrown off by a speaker whose racial features don’t seem to match their speech. “The social networks of the people who are being tested likely matter,” she said.
Future studies could take a closer look at why some individual listeners seem to be affected by the racial features of a speaker. “In the big picture, my lab is interested in understanding and improving speech communication among all kinds of people,” Van Engen said.
Header image: Christina Morillo / Pexels
