The Conversation: Hearing and Speaking
Language Acquisition Before “Mama” and “Dada”
From top to bottom: A three-year-old responds to stimuli; Singh discusses results with Sarah Nestor, left, a graduate research assistant in her lab, and Ashley Yull, a graduate student in psychology; subjects are observed through a one-way window; Nestor administers a test.
Leher Singh, assistant professor of speech, language, and hearing sciences at Boston University's Sargent College of Health and Rehabilitation Sciences, studies language acquisition at its very earliest stages and quickly points out the irony of the way her target population is labeled: “Infant,” she explains, “comes from infans, which is a Latin term for 'without language.'”
Singh has done much to prove the contrary. Her work has demonstrated that babies as young as seven months are able to recognize and remember words they cannot yet say—and that this is an important prerequisite for speech. A recently completed longitudinal study found that children who were better at tracking words as babies had substantially bigger vocabularies at age three.
Clearly, it is not easy to evaluate word recognition in a population that cannot yet do more than burble. However, Singh and her research assistants are able to do it by training a baby to look at a flashing light while listening to a particular word over and over. Then, when the word is used in a sentence, the baby's gaze will be directed toward the light measurably longer if he or she recalls the word.
Singh explains what's most interesting about her results. “We found that differences in vocabulary weren't due to IQ. Demographic factors—parenting styles, whether the children were in day care, household income—had no reliable effects. What was predictable was that if babies did better on our memory tests at seven, nine, or ten months, they had bigger vocabularies at age three. Clearly, we are picking up on something developmental.”
A number of the infants identified in Singh's study as being at risk for language delays were indeed independently flagged by their pediatricians at between 18 and 20 months. Singh points out, “Language delays are not just about vocabulary, but vocabulary is clearly a big part. It's important to identify the risk markers as early as possible, because language acquisition becomes more difficult the older a child is.”
Helping Singh uncover the ways babies master language is undergraduate Chandni Parikh, a human physiology major. “As a sophomore, I wanted research experience,” Chandni explains, “and Leher Singh was very welcoming. Fortunately, I love kids.”
Over the course of the last three years, she assisted Singh with the longitudinal study of infants' word recognition. Thanks to a grant from the Undergraduate Research Opportunities Program (UROP), Chandni was able to spend all of last summer in Singh's lab, helping to finish the study. “It was an amazing experience for an undergraduate. My sister is a senior in high school and has friends applying to BU. I tell them all, you have to do UROP.”
She recently helped Singh conduct a study on affect. Chandni explains, “We played certain recorded words to babies, with either a high or a low affect—in other words, with a happy or neutral tone. Then we gave them the same words, in the context of a sentence, a day later. The children who had heard the words in a happy tone recognized those words far better the next day, even when they were now said in a neutral tone. The children who'd heard those words in a neutral tone couldn't recognize them a day later, even if they were now said in a happy tone.”
“If babies did better on our memory tests at seven, nine, or ten months, they had bigger vocabularies at age three. Clearly, we are picking up on something developmental.”
“It seems that it's better for kids to learn language in 'motherese,'” says Chandni, referring to the slow, musical, hyper-inflected, and emotional style of speech adults tend to use with babies.
Singh is currently interested in seeing whether the lack of such a preference for “motherese”—or other early indicators that language acquisition is not progressing normally—can be used to predict particular disorders. For example, collaborating with the Autism Center at the University of California-Davis Medical Center, she has found that the six-month-old siblings of autistic children—who are at a high risk for autism themselves—tend to prefer more neutral speech to “motherese.” Eventually, Singh's work may allow autism to be identified months or years earlier than it generally is today and allow therapies to be initiated in infancy.
Singh seems to be inspired by the mystery at the heart of human speech. “Language is an incredibly complex system,” she says. “Most of us can't characterize its grammar until we learn a second language in school. Yet we have an implicit knowledge of the system's rules and use them before we can tie our shoelaces.”
She admits that working with a pre-verbal population prone to crankiness when sleepy or hungry can be challenging. “Babies are not sympathetic to the fact that you need the data for some deadline,” she laughs. “But the value of the questions you can ask with them outweighs everything else.”
For more information, see http://people.bu.edu/leher.
