UW News

July 14, 2003

Social interaction plays key role in how infants learn language, studies show

Social interaction apparently plays a far more important role in how infants learn language than previously believed, according to three related studies conducted by researchers at the University of Washington’s Center for Mind, Brain & Learning (CMBL).

The researchers, headed by neuroscientist Patricia Kuhl, reported today that 9-month-old American infants who listened for less than five hours to Mandarin Chinese spoken by native-Mandarin speakers were able to distinguish phonetic elements of that language. In a second study, another group of American infants heard the same material watching a professionally produced DVD or listening to an audiotape but could not distinguish phonetic units of that language.

The findings being published today in the Proceedings of the National Academy of Sciences (PNAS), are the first experimental demonstration of phonetic learning from natural exposure to language under controlled laboratory conditions, according to Kuhl, who is the co-director of CMBL and a UW professor of speech and hearing sciences.

“These findings indicate that infants can glean phonetic information from first-time foreign-language exposure in a relatively short period of time at 9 months of age, but only if the language is produced by a human,” she said. “This suggests social interaction is an important component of language learning.”

In a companion study published in April in the journal Developmental Science the researchers examined the impact of the quality of mothers’ speech on infants’ language acquisition. They found the first evidence that supports the idea that it’s not what you say but how you say it.

“There is an extraordinarily high association between how clearly a mother speaks and how well her baby distinguishes speech sounds, and this is an ability critical to language learning,” said Kuhl.

In this study, Kuhl and Huei-Mei Liu, her former doctoral student who is now an assistant professor of special education at National Taiwan Normal University in Taipei, analyzed the quality of a language called “motherese.” This “baby talk” is used almost universally by mothers and other adults in talking to babies and is characterized by a higher pitch, slower tempo and exaggerated intonation compared to language directed toward adults.

Kuhl has shown in earlier work that children are born “citizens of the world,” able to distinguish among sounds used in all languages. But sometime in the second six months of life, infants begin to concentrate on learning their native language sounds and lose their ability to distinguish sounds important to foreign languages. That is why many adults have difficulty learning a foreign language.

In two studies reported in PNAS, infants were tested to see if they could distinguish between two Mandarin sounds that do not occur in English. Americans often hear both sounds as “chee” or “she” These sounds are difficult for adult Americans to distinguish between but easy for native Mandarin speakers.

In the first study, normally developing 9-month-olds were exposed to Mandarin during a dozen 25-minute sessions spaced over four weeks. During these sessions, native Mandarin speakers read from children’s books and played with toys while speaking Mandarin. Four different speakers, two men and two women, conducted the sessions, so the babies heard a variety of speaking styles. A control group was exposed to the same procedure in English.

Both groups then were tested for their ability to distinguish between the two Mandarin sounds using a head-turn conditioning procedure frequently used to test infant speech perception. The infants exposed to Mandarin were significantly better at distinguishing the two target sounds than were infants who only heard English. In fact, the performance of the American infants exposed to Mandarin for the first time was statistically equivalent to infants in Taiwan who had listened to Mandarin for 10 months, according to Kuhl. The results show that the decline in foreign-language speech perception can be reversed with short-term exposure, she said.

In addition, the phonetic learning of Mandarin appears to be long lasting. The American infants were tested from two to 12 days after their last exposure to Mandarin and the researchers found there were no significant differences in their ability to discriminate between the sounds.

“In previous learning studies babies were exposed to artificial languages for a few minutes and no one expected that kind of exposure to produce long-lasting effects,” said Kuhl. “We predicted that learning in this natural situation would produce a longer-lasting effect. We were surprised to see some of our babies hang on to the information for 12 days.”

The second study was similar to the initial study except that half the infants were exposed to Mandarin by a DVD showing the same Mandarin speakers and materials on a 17-inch television. The others got their Mandarin exposure from an audio-only presentation of the DVD.

Later the infants were tested using the head-turn procedure. Results showed that DVD or audiotape exposure did not lead to phonetic learning, Kuhl said. The infants in this experiment scored at the same level as the English-only babies in the first study who were not exposed to any Mandarin. The infants who watched the DVD or listened to the audiotape also paid significantly less attention than the babies who were in the live Mandarin and English conditions.

“Video plus audio or audio-only presentation did not work for infants 9 and 10 months of age,” Kuhl said. “That’s not how infants learn language.”

She added, “Babies are very sophisticated language learners who use every clue provided to learn — the sounds they hear, their statistical distribution and the social clues provided by speakers — to crack code. The babies were mesmerized by the sight and sound of the foreign language speakers. You could see their little brains absorbing the information.”

Participants in the study published in Developmental Science were 32 pairs of Mandarin Chinese-speaking mothers and their infants. Half the babies were 6 to 8 months old and the others were 10 to 12 months old. Each mother was recorded speaking separately to her baby and to an adult. Later, the mothers’ speech was measured to examine how they expanded the acoustics of their vowels while talking to their babies, a feature previously shown to be related to general speech intelligibility. Finally, the infants were tested for their ability to distinguish three Mandarin vowel sounds using the head-turn procedure.

“Mothers with the highest speech clarity tend to have children with the highest speech perception,” said Kuhl, adding that all mothers modified their speech when talking to their babies, exaggerating the way they articulated vowels.

“All of this is very subtle. We are talking about millisecond differences in timing and small difference in frequency. What is going on is the same thing that happens when you are speaking to a foreigner. You speak more slowly and clearly because you know they have difficulty understanding you. We do the same thing for infants.”

Kuhl added that results of the study could not be attributed to socioeconomic factors such as parental education, income or jobs, which were controlled for.

“‘Motherese’ sounds silly, but in the second half of the first year of life the brain is in a state of neural readiness for language. We think evolution has prepared the child’s brain for this kind of information at this time in development. ‘Motherese’ is a wonderful, rich and deep language that guides the brain along the path to language acquisition,” she said.

Co-author of both papers is Feng-Ming Tsao, who earned his doctorate at the UW and is now a post-doctoral researcher at the Center for Mind, Brain & Learning. The research was funded by the National Institute of Child Health and Human Development and the Human Frontiers Science Program. 

For more information contact Kuhl at (206) 685-1921 or pkkuhl@u.washington.edu.