There are many ways to sound Black, as Ciku Theuri has discovered
This is a transcript of a Subtitle podcast episode
Kavita Pillay: When you think of a “Black voice”, what comes to mind?
Barack Obama: we have never been just a collection of individuals or a collection of red states and blue states, we are and always will be the United States of America
Kavita Pillay: Or maybe a voice like this?
Keisha Lance Bottoms: I am a mother. I am a mother to four black children in America.
Kavita Pillay: What about someone who sounds like this?
Afua Hirsch: When I was growing up, it was very rare to see people who look like me. I’m from Wimbledon. I was one of very few black children in my school.
Kavita Pillay: This is part three of our series about speech — what we hear, and what we think we hear. How our perceptions can take us down the wrong path. The stories we believe about different kinds of voices, and the ones we tell about our own.
Today, Ciku Theuri takes us on a linguistic journey of Blackness and identity.
Ciku Theuri: I’m a first generation American born of immigrant parents from Kenya. My mom and dad both moved here to receive an education. They’ve been able to achieve their version of the American dream as professors.
My mom teaches Food and Nutrition. My dad teaches Accounting. And of course, public speaking and interpersonal communication is crucial in their professions.
And my parents have been doing this for decades, but my mom had some challenges at first.
Serah Theuri: When I came to the US, people here speak with a lot of power, with a lot of loud voice. I was used to speaking low.
Ciku Theuri: She spoke too softly for Americans, specifically for the Americans she met at her first teaching job at Northern Kentucky University.
Serah Theuri: And so they say, “What did you say?” I hate that!
Ciku Theuri: My mom’s laughing now but she really did hate being told to speak up. As the years went by, her Kenyan accent also evolved. So much so that when she returns to Kenya, people there notice she sounds different.
Serah Theuri: I don’t like that because it makes me stand out. If you’re sounding like an American or your accent is like British, they equate that with, “Oh this person has traveled or has lived among Americans or British so they must be of a different class. They must have more money.” So if you go to the marketplace and open your mouth to purchase something, once they hear your accent your price goes up.
Ciku Theuri: And for that reason, whenever I’m visiting family in Kenya and we’re at a market… I keep my mouth shut. But when people there do hear me speak, they automatically revoke my Kenyan card say I sound like:
Martha Migui: The whites…or wazungus as we call them here.
Ciku Theuri: That’s my cousin Martha.
Martha Migui: My first impression of Ciku’s accent when she spoke was very impressive and very different from how we pronounce words here in Kenya.
Ciku Theuri: She’s probably the first person I’ve ever heard describe my accent in that way. And we struggled to communicate at times…especially when I would say particular words
Martha Migui: I could not understand the words until I asked my sister who was at the university to come, listen and interpret.
Ciku Theuri: More like interpret culturally speaking. These American words I was saying, they had no currency in Kenya.
Ciku Theuri: Growing up in the U.S gave me a dual sense of identity. I was Kenyan at home, eating food like ugali, chapati, wearing kitinges and occasionally mixing my mother tongue, Kikuyu, with English. But when I left the house, I was American. And like my mother, I too stuck out like a sore thumb.
First, I was one of a handful of black kids.
Second, I had to help out with the pronunciation of my name. Again and again.
But when it came to the way I spoke, I fit right in.
The Cincinnati neighborhood I grew up in was white and suburban. We would occasionally mingle with the few other African families in the area. The only interaction I had with American kids who looked like me was at church. The Black kids there expected me to speak something like this:
Nene Leeks: I won this bronze damn olympic media in Beijing, China, for individual synchronized swimming.I bet you didn’t know there was one a thing as individual synchronized swimming!
Ciku Theuri: And honestly, I’ll never sound like Nene Leeks in Glee. Even though the show is set in Ohio.
Nene Leeks: Imma say one thing to you and I only going to say it once. If you pee in my pool, I will kill you!
Ciku Theuri: For some reason, the white kids never flinched or made comments about the way I spoke. Maybe because I sounded like them.
But the black kids? They were very vocal about it.
Allison Johnson: Like you’re an Oreo.
Ciku Theuri: My college friend, Allison Johnson says right from when I opened my mouth, the others thought of me as different. Like I was one of those…
Allison Johnson: Black people who talk proper.
Ciku Theuri: Wait, what?
Allison Johnson: They kind of think of you as an Oreo. Like basically black on outside but white on the inside just because you talk proper. Your tone of voice, certain word choices.
Ciku Theuri: It wasn’t like I was especially proud of the way I spoke. Of course, I was proud to be Black and Kenyan but in some ways, I was unsure of my linguistic identity.
During my high school years, I attended an international boarding school in Kenya. There were plenty of people from all over the world, but I was the Black American who sounded white. I was the Oreo.
After high school I returned to the U.S.
I attended a historically black university in Alabama. It was sprawling with Jamaican accents, French accents, southern accents, you name it. And suddenly my very narrow understanding of Blackness was challenged. I was forced to reckon with this stereotypical presumption I had about the way Black people spoke. It’s a stereotype that’s been perpetuated in this country for centuries.
Nicole Holliday: Truthfully in the history of the United States, most black people have been isolated in black communities and most white people have been isolated in white communities. And so the communities diverged in their ways of speaking.
Ciku Theuri: This is Nicole Holliday. I needed to ask an expert about my speech, and Nicole’s an expert! A professor of linguistics at the University of Pennsylvania. She says it’s hardly surprising that I sound the way I do.
Nicole Holliday: If you sounded really what people stereotype as an African American speaker is it would be weird, like how did you come out that way?
Ciku Theuri: Yeah, that’s true! Nicole also says the fact that I’m young and a woman who grew up in a middle class area has a heavy influence on the way I speak. It makes it harder for people to identify my race without seeing the color of my skin. She also says it’s typical that people like me — kids of immigrants — don’t interact all that much with people who speak what’s considered Black American English.
I ask her, what is Black American English, anyway? Nicole says it all goes back to the fact that Black people didn’t speak English when they were brought to America.
Nicole Holliday: In the conditions of slavery where they were acquiring English in this unusual way — they weren’t being taught it in schools — and they were getting input from all these different Europeans, and they had their first language. So their first languages were being influenced by the English that they were speaking. So the way that Black people were speaking in the United States has always been different from the way that most white people were speaking.
Ciku Theuri: And there’s a lot of overlap between Black American English and southern US English. Most enslaved Africans lived in the south. But later, Black Americans started moving elsewhere across the country. This was the time of the Great Migration, between 1916 and 1970. And of course the way they spoke evolved. No wonder people with those backgrounds speak differently than the way I do.
Nicole Holliday:…and really, like, we are carrying around our ancestral experiences as well as our personal experiences in the language that we speak. And that’s something that you’re getting at here.
Ciku Theuri: The dimensions of Blackness are something I learned to embrace and celebrate once I started college. I came to understand that Blackness meant more than American Blackness. I realized it included me, the child of immigrant parents. I met other students who grew up in American suburbs and learned to accept the fact that I didn’t have to speak a certain way to be considered Black. We were Black enough. I was Black enough.
Nicole Holliday: It doesn’t mean that it’s a problem that you don’t sound the way that people expect you to. It just means that you had a different experience.
Ciku Theuri: For a long time, I was unsure of how burdensome or beneficial the way I spoke would be for my future. Would it advance me professionally but isolate me socially? Would I only be accepted in certain circles and excluded from others? But after college I realized that the weight of my linguistic identity may not be as significant after all.
Nicole Holliday: It’s not an accomplishment to speak African-American English, and it’s not an accomplishment to speak mainstream English. People just talk the way they talk as a function of their experiences, and we should really be OK with all of that.
Ciku Theuri: And maybe people are beginning to believe that notion. Maybe they are becoming more open to the idea that Black people are not a monolith …that we don’t all speak the same way. I think we’re seeing a little bit of that now, as the Black Lives Matter movement continues its momentum, there are more calls for the uplifting of Black voices to share their beliefs… their experiences…but yet, when this happens, I still hear people I know react by saying, “Oh, they’re so articulate!” Or “Wow, they speak so well!”
Nicole Holliday: What you’ll hear people say sometimes is things like, “oh you think because I’m Black that I should sound ‘ghetto?’” Well, this idea that there’s a problem with sounding identifiably African-American is just racism, it has nothing to do with the language itself.
If they have that preconceived prejudice, it doesn’t matter how you say things or what you say: they’re always going to complain.
Ciku Theuri: My mom, like me, has learned to embrace her own accent.
She’s still teaching in the Midwest, but now no longer conflicted about the way she speaks.
She kicks off each semester saying this to her new students:
Serah Theuri: My accent is different. I’m not from Southern Indiana. I know that whoever listens to me speak understands what I’m saying. That’s when I realize that it’s an asset. It’s what makes me me.
Ciku Theuri: I have to agree with her. The distinct way in which I speak….different from my mom AND different from many other Black Americans…is what makes me me.
Kavita Pillay: I keep thinking about this thing that Nicole Holliday said.
Nicole Holliday: We are carrying around our ancestral experiences as well as our personal experiences in the language that we speak
Kavita Pillay: We’re carrying around the voices of the people who came before us. What does that mean for the future, a future with more Siris and Alexas? Siri, what does it mean now?
Siri: You’ve got me!
Kavita Pillay: Well, in April, Stanford came out with a study showing that voice recognition tools by the tech giants — including Apple, Amazon, and Google — misidentified words spoken by Black people at a far higher rate than those said by white people. Which means that Black Americans could face even more hurdles when speech recognition is used in other settings. Like a courtroom or the workplace. And Black people are less likely to benefit from a voice assistant, like Siri or Alexa.
But Siri and Alexa aren’t the only voice assistants in town.
Morgan Freeman: Your toast is ready.
Kavita Pillay: Morgan Freeman is now the voice behind Jarvis, which is Mark Zuckerberg’s personal voice assistant. Issa Rae has also been a celebrity voice assistant for Google.
Issa Rae: And right now I’m using Issa’s voice, which is pronounced EE-sa, not Ih-suh
Kavita Pillay: And before her, it was John Legend.
John Legend: He would probably tell you he don’t want to brag, but he’ll be the best assistant you ever had.
Kavita Pillay: Samuel L. Jackson was the first celebrity voice for Amazon. By the way, his voice came in PG and R-rated versions, in case you didn’t want him telling your kids to Go the [Bleep] to Sleep
Samuel L. Jackson: Hello boys and girls, it’s been a long day and you are due for some sleepy times!
Kavita Pillay: Morgan Freeman. Issa Rae. John Legend. Samuel L. Jackson.
There’s a few common denominators here: the tech giants are battling hard to get voice assistants into our lives. And recruiting celebrities to lend their voices makes a voice assistant an easier sell. I’m more likely to forget that Google is collecting all kinds of info on me when Issa Rae is at my beck and call, making me laugh. Hey tiger!
Issa Rae: You can still call me your Google assistant. Now I just sound extra fly!
Kavita Pillay: Is it just a coincidence that all of these celebrity assistants are Black? Maybe. But how about this: in 2019, Adobe surveyed a thousand digital device owners on the celebrity voices that they would like as an option. And four out of the top six — The Rock, Will Smith, Chris Rock, and Oprah — are also…Black. Or in the case of The Rock, half Black, half Samoan.
Yes, they’re all household names. They all have distinct voices. And yes, a digital assistant is sort of a tech demigod, all knowing and omnipresent. But we don’t call them demigods; we call them…assistants. Why do so many people want the world’s most famous Black entertainers to “assist” them, to serve them, silent until spoken to?
The tech giants seem to be doing a great job of commodifying Black voices, at least a handful of them. But if you’re a Black person who owns a smart speaker, the minute you deviate from a “white” speech pattern, the voice recognition software gets more and more clueless.
Sundar Pichai: Our goal is one day to get the right accents, languages and dialects right globally.
Kavita Pillay: Get the right accents, languages, and dialects…one day. Sundar Pichai isthe head of Google. So what’s the problem now? Why hasn’t it happened yet?
Ian McLoughlin: The technology is based around big data.
Ian McLoughlin is a professor at the Singapore Institute of Technology. His research area is artificial intelligence for speech.
Ian McLoughlin: So anywhere where we see the words artificial intelligence — that’s backed up by a big data approach. Which means large amounts of recordings. These recordings are needed to make models. And models require five things:
Ian McLoughlin: Big data scientists talk about the five Vs that they require in their databases.
Kavita Pillay: Volume.
Ian McLoughlin: Velocity.
Kavita Pillay: Variety.
Ian McLoughlin: Veracity.
Kavita Pillay: And value.
That first one — volume — is really important. Volume is the foundation of big data. And it helps explain why the current systems are not adept at interpreting a variety of accents.
Ian McLoughlin: There just isn’t the volume of language. There isn’t enough recording there to get a good statistical grasp of the features in their speech. If you can’t get a good statistical grasp, you can’t model it.
Kavita Pillay:: The data also needs to be updated frequently, which is the second ‘v’, velocity. It has to take a lot of different forms. That’s variety. And it needs to be of good quality, which is the fourth ‘v’, veracity. And value? That’s the ability to take this ocean of data and turn it into a successful business.
After all, these are companies that are building products that learn. The more we use them, the more they learn. And companies have little incentive to build products for populations that speak smaller languages, or that do not have the dominant accent.
Remember the research that found voice assistants are more likely to misidentify words spoken by Black Americans than those who are white? It comes down to this: the data used to train these voice recognition systems comes largely from white people.
This matters more than ever because voice activated technology is becoming omnipresent. And that’s frightening for a few reasons.
But when it comes to language, we’ve been here before. We’ve been scared about what technology might do to the way we speak.
Ian Millar: I remember when the World Wide Web first came into being in 1994, almost everything that was on the World Wide Web was either an English or Japanese at that point, but largely English.
Kavita Pillay: Robert Millar is a Professor in Linguistics and Scottish Language at the University of Aberdeen in Scotland.
Ian Millar: People said this is terrible for all the other languages, particularly the small scale ones. But as it turned out, it’s actually enacted the liberation of these languages quite often because you don’t need to be in Brittany to speak Breton anymore.
Kavita Pillay: That turned out to be true for languages. So what about accents? Robert Millar is a Scot, and as it turns out, the Scottish accent is one that voice recognition technology has trouble understanding…
Comedy sketch: Voice recognition technology? In a lift? In Scotland?
Kavita Pillay:…As in this super popular British comedy sketch about a lift, an elevator, with a voice assistant.
Comedy sketch: You ever tried voice recognition technology? They don’t do Scottish accents.
Kavita Pillay: The sketch ends with the frustrated characters still stuck in an elevator, rallying for Scottish independence.
Comedy sketch: [Automated voice] You have not selected a floor [Actor 1] Up yours, you cow! Scotland ya bastard! [Actor 2] Scotlaaaaaand! Freedoooooooom!…
Kavita Pillay: This sketch happens to be from the BBC. And the BBC seems to be paying attention to the call for voice assistants with more nuanced accents.
Beeb: I’m new, so I can’t do everything.
Kavita Pillay: That’s Beeb, the BBC’s voice assistant which is currently in beta.
Beeb: But, you can ask me to play any of the BBC’s radio stations, shows, and music mixes.
Kavita Pillay: Beeb doesn’t have the classic British accent or what’s known as Standard BBC English. It’s more of a generic northern British accent. And at least to British ears, this accent is considered warm and friendly. But, it’s not really an accent that anyone speaks with in Britain.
Beeb: Why not give it a try by saying, OK Beeb, update me.
Kavita Pillay: Whether we know it or not, you and I are carrying around the voices of the past. Are voice assistants revealing how we might speak in the future? They know a lot about us already, but when it comes to how we speak, our voice assistants still have a lot to learn. Hey Google, and Siri, and Alexa, if you’re listening? The stakes are too high for you to screw this up.
Siri: Hmm. Is there something else I can help with?
This episode of Subtitle was reported by Ciku Theuri, with additional reporting by Kavita Pillay and Lydia Emmanouilidou.
Subtitle’s sound designer is Tina Tobey.
This episode was edited by Julia Barton and Patrick Cox.
Thanks to: Richard Bell, Lauren Daly, Mukul Devichand, Jeremy Helton, our partners at the Linguistic Society of America and the Hub & Spoke audio collective.
Subtitle is funded by the National Endowment for the Humanities.