Accents Are Forever

By their first birthday, babies are getting locked into the sounds of the language they hear spoken

The 6-month-old lies happily in the crib, cooing and babbling to the beaming parents and grandparents who respond encouragingly, eagerly anticipating that first word, that distinguishable "Mama" or "Dada." But, of course, 6 months is too soon to expect such a landmark utterance. Not until somewhere around the first birthday will Baby have developed the power of speech and thus be able to repeat and imitate those cute sounds the adults are making. Prior to that, words are just background noise that doesn't register on the infant brain. Right?

Sorry, wrong, Patricia Kuhl is saying. Well before that first word is joyfully recorded on Baby's growth chart, according to Kuhl, who is director of the Center for Mind, Brain, and Learning at the University of Washington, the supine little person is taking in and carefully filing for future use the nuances of pronunciation—and mispronunciation—of the language heard at cribside. As early as 6 months, Baby begins to differentiate one bit of verbal input from another, mentally reviewing and silently rehearsing for the moment when the vocal cords are ready to deliver. And when the words do tumble out, they will be reproduced with the distinctive, localized twang or lilt overheard in Mommy's and Daddy's (or Mama's and Papa's) own funny speech. Accents, in whatever language, stubbornly hang in there for years, decades, a lifetime, without being easily rubbed out. Aha, I said, thinking of my Southern mother, who said "heah" for "here" until her dying day, my own persistent Pennsylvanianisms, and my wife, Sally.

She was born in the Philippines, began speaking English at 6, has lived in English-speaking Hong Kong and the United States for more than 30 years, holds a doctorate in education from Stanford, and still cannot easily twist her tongue around the English tz and ch sounds, which blend together in the Philippine languages she learned as an infant. Thus we jokingly snack on "Rich crackers," and have a friend named "Dorothy Ritz." She replies that former Pittsburghers stress the first syllable of "Hello"—"Hell-low."

"Henry Kissinger wasn't born with an accent," Kuhl said, of the German-born ex-Secretary of State who often joked about his guttural pronunciation, "and neither was your wife. He speaks English beautifully, and I'm sure so does she. Yet you can tell they're not native speakers of the language. Why is that? Our research shows that a kernel of that pattern of speaking begins to form in the brain well before actual production of speech. And by the time the baby's first words do come, those distinctive characteristics are solidly in place."

For 25 years, in her laboratory overlooking Lake Washington, Pat Kuhl has been exploring how humans develop spoken language and why, for example, American English speakers can readily distinguish between the l and r sounds and adult Japanese are consistently defeated by "lake" and "rake." Or, for that matter, why Americans struggle to hear the subtle difference between the b and p sounds in Spanish, so that "beso" (kiss) becomes "peso" (weight). Working with American, Swedish, Japanese and Russian children, Kuhl has discovered that 6- to 8-month-olds clearly hear and respond to the vowel and consonant sounds in both their own and other languages; 6-month-olds in Tokyo perceive the difference between l and r just as easily as babies in Seattle. But by the time the same infants are a year old, they have lost that ability. Instead, they zero in on the "home" sounds and tune out unfamiliar ones. The first language keeps fighting off the pronunciations of a new language, so that Kissinger indelibly says "Vashington," my wife says "Dorothy Ritz" and the Chinese students in Kuhl's lab double up with laughter as American colleagues learning Mandarin mix up "shee," meaning "west," with "chee," or "wife."

How in the world do we know that sounds in any language register with a 6-month-old, who can't even say "Dada" yet? Kuhl leads me into an adjoining room equipped with table, chairs, a small loudspeaker and an odd-looking plastic box. She waves me into a chair. "We call it the 'head-turn' study," she says in a voice that is decidedly Minnesotan. "Mom sits here at the table, with Baby in her lap. An assistant sits across the table, moving a toy that keeps the baby's attention. From the speaker comes 'La la la la la' at one-second intervals. Baby is listening, but watching the toy. Then the sound changes to something different: 'La la ra.' And when that happens..."

The plastic box lights up. A teddy bear begins to dance and beat vigorously on a drum. "Baby learns that when the sound changes, the bear performs. They turn to watch and get this fancy reward. At 6 months, two out of three Japanese and American kids turn to look. At a year, 80 percent of Americans turn, but only 59 percent of Japanese, just a little better than chance. To them, the two sounds have become the same."

The light goes out and the bear puts down his drumsticks. "Think of it like a computer," Kuhl says to explain how a tiny infant can register sounds and then reproduce them months later. "You feed in data and more data, and it's there in the computer, and one day you hook up a printer here"—she points to her mouth—"and it comes right out."

Back in the mid-'70s, when she was a postdoctoral scholar at the Central Institute for the Deaf in St. Louis, Kuhl showed that monkeys responded to sound changes in the same way, although of course they could not produce recognizable speech. Her report was greeted with polite skepticism. Not at all fazed, when she came to Seattle, she adapted an existing technique for testing infant hearing into the head-turn experiment with human babies. More skepticism: conventional wisdom held that infants' brains weren't sufficiently developed to process sounds and wouldn't be until they were old enough to imitate adult speech on their own. Gradually, Kuhl's has become the accepted view, to the extent that she was recently featured at a White House conference on development of the infant brain.

A baby's brain, Kuhl says, is a work in progress. Even before birth the brain's zillions of neurons, or nerve cells, are reaching out to each other to make connections, or synapses, forming the intricate wiring that guides all life. The pathway from hearing sound to interpreting its meaning is one example.

"The baby early begins to draw a kind of map of the sounds he hears," Kuhl says. "That map continues to develop and strengthen as the sounds are repeated. The sounds not heard, the synapses not used, are bypassed and pruned from the brain's network. Eventually the sounds and accent of the language become automatic. You don't think about it, like walking. The sounds your wife heard earlier become more and more embedded into the map, until eventually they are almost ineradicable."

Moreover, sounds are sorted by what Kuhl calls "a magnet effect." Those resembling a prototypical sound in the home language are taken in and interpreted as the home sound. Between the magnet effect and the detailed mapmaking, any attempt to introduce a new language creates "interference," in Kuhl's term. With each passing year, redrawing the map becomes more difficult.

But, I asked, what about children who grow up bilingual? All of us have known kids who were fluent in English on the school playground but spoke another language at home. In fact, Sally has been equally facile from childhood in two decidedly different languages—Cebuano, the language of the central Philippines, her mother's home, and Tagalog, her father's tongue, which in its standardized form is the national language spoken around Manila. In those cases, the infant brain simply draws two maps, and the process is particularly easy when a specific language can be identified with the tone, pitch and pronunciation of each parent.

Not that an adult absolutely can't learn another language, cautions Kuhl, who herself is trying valiantly to master Mandarin Chinese. None of us want to believe that our brains are inexorably fixed, and we can't learn new tricks as grown-ups. But becoming fluent and accent-free in a new language becomes increasingly difficult, and the best time to start is as early as possible, like nursery school, while the brain is still developing. By puberty it can be an uphill struggle, as generations of language students have found. "People talk about a 'window of opportunity' for learning language," Kuhl says. "The implication is that if you miss that opportunity, it's too late. I don't agree. It is more difficult with the years. But not impossible."

English is rapidly becoming the language of science and of computers. In Silicon Valley, just down the road from where I live, you hear the accents of Bangalore, Helsinki, Tehran and Taipei randomly mixed in with pure Yankee. The newspapers carry ads for "accent reduction clinics," and as many as half the kids in a classroom may speak another language at home. All of which makes Kuhl's research about early language learning particularly relevant.

When I arrived home from visiting Kuhl, Sally was waiting for me. "Hell-low," I said. "You ought to do something about that accent," she replied.

Get the latest Science stories in your inbox.