The Ethics of The Accent-Reduction Industry
This is what's missing from the conversation on accent reduction and accent discrimination.
Accent discrimination is as old as discrimination itself. One of the earliest recorded examples is in the Bible. In the eleventh century B.C, two Semitic tribes—the Gileadites and the Ephraimites—were separated by the Jordan river. Separation of people meant a separation of language. And over time, the language of the tribes morphed into similar dialects.
When a war broke out between the two tribes, the Gileadites won. The Ephraimites then tried to flee home, but the Gileadites blocked them at the river. At this point, the Gileadites had to find a way to identify members of their tribe and members of the enemy’s tribe.
To do that, they used language.
The word for ‘ear of grain’ was pronounced differently by the two groups. The Gileadites pronounced the word as shibboleth; the Ephraimites pronounced the word sibboleth.
The minor change of sh → s, a mere matter of moving the tongue slightly forward, led to the Ephraimites being killed as soon as their true identity was revealed.
Identifying the enemy through language has been used throughout history in different wars. The Dominican military under Trujillo used the word perejil to identify and kill Haitians. The U.S. military during World War II used the word lollopalooza to identify Japanese spies.
Accents reveal identity. And identity can have harmful consequences. Outside of the context of war, having an accent could lead to having less access to quality jobs, it can lead to being fired from jobs1, it can lead to a child being misdiagnosed with a speech disorder.
The fear of having the “wrong” accent is valid; the yearning for the accent with the most opportunities attached is valid.
But is the fear and yearning being assuaged by the completely unregulated multimillion dollar accent reduction industry? In the industry, anyone can claim to be an accent reduction coach. And oftentimes, the courses offered are hundreds or even thousands of dollars.
This week, I don’t want to dissect the ethics of particular accent reduction programs. Instead, I want to add to the conversation pieces of information that are important to know before signing up for or offering accent reduction classes.
Accents Recap
For the past two weeks, we’ve looked at pronunciation and accent. We’ve made an important distinction between comprehensibility, intelligibility, and accentedness. Here are the definitions one more time:
comprehensibility: how difficult you are to be understood
intelligibility: how much of your message is understood
accentedness: how “different” you sound from the “norm”
All three of these dimensions are separate. Research has shown that quality pronunciation instruction can absolutely help speakers with comprehensibility and intelligibility in their second language. As for accentedness, research has shown that it’s hard--if not impossible--to completely sound like a “native” speaker in your second language2. That being said, quality accent reduction instruction can make you sound more “native” like.
For the past two newsletters, I used research from phonetics, phonology, and second language acquisition. Those lines of research are really helpful because they tell us how individuals can improve their pronunciation based on the latest understanding of how language works in our minds and bodies. But a drawback of that research is that it takes language out of a social context—it focuses on what the speaker can do to improve their pronunciation and it treats all speakers the same way.
In a Social Context, Every Conversation Has a Speaker AND a Listener
Putting language inside of a social context acknowledges that there are (at least) two people in any given conversation: a speaker and a listener. And linguist Lippi-Green says that the social space between the speaker and listener is rarely neutral.
When a speaker has an accent the listener is not familiar with (or is biased against), the listener has two choices:
They can make the effort to listen a little more closely and have a conversation.
They can say, “I don’t understand you,” and end the conversation there.
Lippi-Green says that whether the listener chooses option 1 or 2 depends more on the listener’s social evaluation of the accent than on how easy the speaker is to understand.
Let’s take a deeper look at examples of why the degree of accent may not be relevant in the listener’s decision to choose option 1 or 2.
Complicating Accent
If we look at our definition of accent—a different way of speaking—then it’s clear that accents are relative to the listener. And listeners—without always meaning to—are biased. And their biases are not always based on language.
One study from the 1990s3 presented participants with the same recording and a picture of the speaker. Both groups heard the exact same recording. But in one group, they were presented a picture of a Caucasian man as the speaker. In the other group, they were presented a picture of an Asian man as the speaker.
Even though the recording was identical, the participants only heard an accent when the voice was coming out of an Asian man. Even crazier, the participants found the Asian speaker more difficult to understand than the Caucasian man.
Complicating Intelligibility
Intelligibility is also not objective. Listeners use other information, like accent or nationality, to make judgments based on how easy the speaker is to be understood.
In one study4, they made participants listen to English recordings of speakers from the U.S, Italy, and South Korea. The participants were not told where the speaker was from. They just had to listen to the recording and then rewrite words the way they heard them. For example, if the word was “methods,” the participants could rewrite it as “mesods” or “meteds.”
All speakers in the recordings used reduced vowels appropriately. But they were not all perceived the same way when they used reduced vowels. When the American used a reduced vowel, it was not perceived as a mistake. When the Italian speaker used a reduced vowel, it was rewritten more times than the American speaker. When the Korean speaker used a reduced vowel, it was rewritten more times than the American or Italian speaker.
All speakers were doing the exact same thing with their vocal tract. But they were all perceived differently.
Are We Still Talking About Language?!
It’s an important question: Is accent discrimination about language?
To help us think about this, we can use a raciolinguistics perspective which states that you cannot separate language from race or race from language. One of the authors of a raciolinguistics perspective has a book called Looking Like a Language, Sounding Like a Race, and the title does a good job of showing why race and language are so intertwined:
A speaker can look like a language. In the study with the photos of the speakers, the Asian man looked like accented English and the White man looked like “native” English. The language was the same, but the biases the listeners had caused them to create an accent for the Asian man.
A speaker can also sound like a race. This means that the listener could assume the speaker’s race based on their language alone.
If race and language are intertwined, that means that changing your accent might have zero effect on a biased listener. To give Lippi-Green the last word on this:
“The degree of accent is not necessarily relevant…where no accent exists, stereotype and discrimination can sometimes manufacture one in the mind of the listener,” (p. 251).
Conclusion: You Are A Listener
The information I’ve presented today may make it seem like accent reduction is futile and accent discrimination is inevitable, but that’s not the message I’m trying to convey. The message is that when you’re languaging, you’re sometimes the speaker and sometimes the listener. As the speaker, you can’t change how the listener perceives your language. As the listener, you can change the way you perceive people’s language.
Each and every one of us has certain ideas about language. When it comes to English, we have ideas about which Englishes are better than others, which dialects we consider “broken English” or “uneducated English,” which accents we hear as accented (and what messages we attach to the accent).
We don’t need to be exposed to a larger diversity of accents. We need to become aware of what thoughts come to our minds when we hear certain accents and why those particular thoughts are coming up.
This is true when we’re listening to others’ accents.
This is extra true when we’re listening to our own accent.
Works Cited
Lippi-Green, R. (2012). English with an accent: Language, ideology, and discrimination in the United States. Routledge.
Rosa, J. (2019). Looking like a language, sounding like a race. Oxf Studies in Anthropology of.
(The rest are in the footnotes)
This is illegal in the U.S. It has happened in the past, though.
“There is no known strategy that can eliminate foreign accents in adult learners,” (Derwing & Munro, 2014, p. 73)
Rubin, D. L. (1992). Nonlanguage factors affecting undergraduates' judgments of nonnative English-speaking teaching assistants. Research in Higher Education, 33(4), 511-531.
Lindemann, S. (2017). Variation or ‘error’? Perception of pronunciation variation and implications for assessment. Second Language Pronunciation Assessment,, 193.