qiy's linguistics world

A viral video from 2013 went like this: a White man approaches an Asian American woman and strikes up conversation. After a few exchanges, he asks, “Where are you from? Your English is perfect.” – a question that has become cliché to many Asian Americans. She responds, “San Diego. We speak English there.” Unsatisfied with this answer, he proceeded to ask a series of…embarrassing questions about her ancestry.

The 10 million+ viewers can sit back and laugh at this bumbling, insensitive man for making a fool of himself. But the video also hits on an underlying problem not exclusive to internet caricatures: an assumption of “foreignness” when faced with an Asian speaker – even when her “English is perfect.”

The phenomenon, often called reverse linguistic stereotyping, is where listeners ascribe features to a person’s speech based on their appearance or aspects of their identity. For the man in the video, hearing native English from an Asian woman was surprising. In other cases of reverse linguistic stereotyping, listeners believe an Asian speaker has a non-native accent even when they are a native speaker.

Studies on Reverse Linguistic Stereotyping

In 2009, Okim Kang and Donald L. Rubin recruited a diverse cohort of 158 college students and presented them with the following:

• A photo of a White “Anglo” male face introduced as “Mr. X born in the U.S., an American teaching assistant.”
• A photo of an East Asian male face introduced as “Mr. Y born in China, an international teaching assistant.”

Then, the listener is told that Mr. X will give them a lecture on space and galaxies. A four-minute recording of Mr. X’s voice would play. Next, the listener is told that Mr. Y will finish the rest of the lecture, and a recording of Mr. Y’s voice would play. (Some listeners were presented with the opposite condition: they heard Mr. Y’s lecture first, then Mr. X’s.)

Listeners rated Mr. X and Mr. Y’s lectures on teaching quality and “accent standardness.” Listeners were also given transcripts of Mr. X and Mr. Y’s lectures with some blanks, which they were prompted to fill them out.

On average, Mr. Y’s teaching quality was rated lower, and his speech was perceived to be more “accented.” Listeners also filled out less of Mr. Y’s transcript, suggesting poorer listening comprehension. Even listeners with Chinese and East Asian backgrounds displayed these biases.

The catch: unbeknownst to the raters, Mr. X and Mr. Y’s voices were the same voice, as the lectures were produced by the same native speaker of standardized American English. In fact, the speaker was a communications professor known for his “clear speech” and a model of “good English” to many.

The researchers concluded that the association of a Chinese person with lower “prestige” and more “foreignness” caused a perception of non-native speech, and thus lower intelligibility.

Is it “pattern recognition,” or is it prejudice?

Pattern recognition is the instinct that drives us to be outraged when eggs cost $6.00 a dozen (definitely contrary to prior experiences we’ve stored in our heads!). So it may not be that Kang and Rubin’s subjects were bigoted, but rather that they lacked exposure to proficient English speakers who are East Asian. I’m Asian, and I confess I’d be taken aback if a White person spoke Cantonese fluently. And we’d all be taken aback if a child spoke in a gruff, low voice, as that falls contrary to our experiences with children.

However, English occupies a unique place in the global stage, so it is not analogous: it is spoken as a native language all across the world, from Ireland to Singapore to Uganda. A single culture, race, or ethnicity cannot lay claim to languages like English ¹ – and to imply so would have dire consequences.

Perhaps more diverse, more bilingual societies would be less likely to associate English nativeness with Whiteness. Kutlu et al. (2022) found that participants in Montreal, Canada perceived no difference in intelligibility or accentedness between native English speech paired with a White face and the same stimuli paired with a South Asian face. However, in Gainesville, FL – a less bilingual city (and arguably where bilingualism is also not as highly valued) – participants rated more accentedness and transcribed speech less accurately when presented with the South Asian face – a finding similar to Kang and Rubin’s.

Was it pattern recognition, or was it a culture of prejudice? As mentioned above, Gainesville has a lower bilingual population compared to Montreal, so Gainesville residents likely encounter non-native speech less frequently. However, Gainesville is about 50% non-White, and 17% of the city speaks a language other than English at home. That is not insignificant. A diverse cohort of college students was recruited for Kang and Rubin’s study³. If college students, who are relatively liberal-minded, display such biases, then there is a concern about these biases in wider society.

To quote an analysis by Babel (2022), “In parts of North America, stereotypes about ethnicity and accent associations present a socio-linguistic landscape where white individuals are licensed to be native speakers of English, whereas non-white individuals are presumed to be second language speakers of English.”

Linguistic stereotyping and reverse linguistic stereotyping

For decades, studies on linguistic stereotyping have found negative biases against non-native English speakers (Todd, R. W., & Pojanapunya, 2009) and speakers of non-mainstream dialects (e.g., African American English, Parker (2021)). Unfortunately, to combat these biases, speakers of minoritized or non-mainstream dialects may be told to stop speaking their dialect. Some non-native speakers believe they would be viewed more highly if they “had less of an accent.” But studies on reverse linguistic stereotyping suggest something harrowing: that you can completely emulate the voice of a native speaker of Standard American English, and yet some listeners will continue to only hear what they expect to hear.

So what can we do?

I have several ideas on how to combat these biases (e.g., re-examine oral proficiency requirements in academia- and occupation-related English assessments, teach HR departments about linguistic and reverse linguistic stereotyping) – but I think a good start is this: When you continuously ask a non-native speaker to repeat themselves, when you find yourself speaking over a non-native speaker, or when you fall back on blaming someone’s “accent” for your lack of comprehension – examine your biases.

-qiy

---

References

---

1. Point of clarification: A culture or ethnicity can, however, identify with a particular variety of English – such as African American English, commonly spoken by Black Americans; or Singlish, commonly spoken by Singaporeans. ↩︎
2. An additional anecdote: Dr. Kang reported to NPR that she does a similar experiment every semester in her introductory linguistics course, and the results are the same: lower comprehension and greater accentedness rating to an Asian guise with a “native English” voice. ↩︎