By the age of three, wealthy children hear 30 million more words than their poor counterparts, according to a landmark 1995 study (pdf) by researchers Betty Hart and Todd Risley. The “30 million word gap” has since become shorthand for the gaping inequalities between high and low-income children. Hart and Risley also showed that the number of words kids heard by their third birthday strongly predicted kids’ academic success when they were nine. Clearly, it was important to talk to children—a lot.
But a new study from researchers at MIT, Harvard University, and the University of Pennsylvania shows that it may not be the sheer accumulation of words that builds children’s brains and their verbal and non-verbal skills. Rather, “conversational turns,” or back-and-forth banter, proved to be much more predictive of a child’s language development than the number of words spoken to them.
By using brain scanners, natural language processing systems, and a raft of standardized tests on four-, five-, and six-year-olds, the researchers found that conversational turns strongly correlated with both more brain activation and stronger scores on tests. Their findings applied to children regardless of parental income or education.
“The sheer amount of language children heard spoken by adults wasn’t linked to children’s brain responses, but the number of conversational turns was,” says Rachel Romeo, lead author of the study and a doctoral student at a joint Harvard-MIT program. This suggests, she says, that providing children with more opportunities to experience language through educational apps or TV programs isn’t enough. Our focus should be on activating children’s brains through what scientists call “serve and return”—the process by which an infant or child “serves” to an adult, in the form of a gaze, a sound, or a question, and the adult returns the serve with an affectionate and engaging gaze, a coo, or a caring response to that question.
“We worried the message around the word gap was suggesting that parents dump words into their kids,” said Romeo. “The brain response show us that’s not the case; it’s the interactive language, the back-and-forth that is most important for the brain.”
Romeo and her colleagues believe that these conversational turns help to actually rewire and grow kids’ brains. “It’s almost magical how parental conversation appears to influence the biological growth of the brain,” says John Gabrieli, a professor and member of MIT’s McGovern Institute for Brain Research, and the senior author of the study.
The study included 36 Boston-area children from diverse backgrounds, aged four to six. First, researchers tested kids’ verbal and reasoning skills. Then researchers looked at functional magnetic resonance imaging images (fMRI) of the children’s brain activity while they listened to audio stories. Afterward, families were sent home with digital voice recorder small enough to fit inside a child’s pocket, which can record up to 16 hours of language spoken at home.
Algorithms then analyzed the recordings, counting the words spoken by adults and children, as well as conversational turns (when an adult and child speak to one another with no more than a five-second lag between words). This is a more complex language measurement. “Whereas adult words and child utterances are simple linguistic measures, conversational turns incorporate both linguistic information and nonverbal communicative aspects such as temporal contiguity, adult responsiveness, join social attention, and exchange of communicative information,” the study says.
Then the researchers put it all together, comparing the children’s test scores to their brain images and audio recordings. They expected to find the same relationship as Hart and Risley—the more words spoken to children by adults, the higher kids’ scores on standardized tests and the more activation seen in kids’ brain. But what they found surprised them.
The number of adult words and conversational turns were both correlated with children’s scores on standardized tests measuring their skill with language. But the relationship between adult words and test scores was fully explained by socioeconomic differences.
Conversational turns, however, were linked to language scores even after accounting for income and education. High-income children with fewer conversational turns had lower language skills and brain responses, performing worse than their incomes would predict. The reverse was true of low-income kids with more conversational turns.
Conversational turns also correlated with more activity in Broca’s area, a part of the brain associated with speech production and language processing. Overall, Romeo and her colleagues found that for every 11 conversational turns, a child’s verbal test score increased by one point.
Like Hart and Risley, the researchers observed enormous disparities in what happens at home. The number of adult words spoken at a peak hour ranged from 1,953 to 6,991. The number of child utterances ranged from 300 to 1,275, and the number of conversational turns ranged from 86-330.
Romeo notes that correlation is not causation. While more conversation at home predicts better test scores and more activity in the Broca’s region, it is not proven that serve and return is the root cause. And the study was small—smaller even than the Hart and Risley study, which looked at 42 families.
The study underscores another critical point about babies and kids: They are social learners. Infants learn from humans who connect with them and make them feel safe.
Kathy Hirsch Pasek, author of Einstein Never Used Flashcards: How Children Really Learn and Why They Need to Play More and Memorize Less, runs Temple University’s Infant and Child Laboratory.
She explained to Guardian, “We arrive ready to interact with other humans and our culture.” As Guardian writer Alex Beard continues, “The real genius of human babies is not simply that they learn from the environment—other animals can do that. Human babies can understand the people around them and, specifically, interpret their intentions.”
In a set of now-famous experiments, Patricia Kuhl, co-director of the University of Washington Institute for Learning and Brain Sciences, studied infants trying to learn Mandarin. Some learned from a teacher, others from the same teacher on a screen, and another, from an audio recording. The kids with the human teacher were able to learn Mandarin, while the other two groups were not. Hirsch Pasek and Roberta Golinkoff from the University of Delaware further unpacked this relationship in a study that found that kids can learn from video, as long as the person in the video responds to the child as a live person would—such as when kids and adults use Skype. But if it’s a pre-recorded video, kids learn nothing.
“These more social aspects seem to be really important for cognitive development,” Romeo says.
According to scientific consensus, the most important time for language acquisition is a child’s early years, up to age five. For pragmatic reasons, Romeo’s study needed to use children who were a bit older. But Romeo notes that the study’s findings amplify the idea that serve and return is critically important at any age, including when the “serve” is simply a gaze or a gurgle from an infant.
“A conversation with a four- to six-year-old looks different than conversations with infants, but the principles of taking turns are the same,” she says. Those principles include taking the time to listen, wait, and not always provide answers. That’s not easy in our fast-paced times—but it’s worthwhile, considering what’s at stake.
This reporting is part of a series supported by a grant from the Bernard van Leer Foundation.