How babies learn – and why robots can’t compete
I asked her about language learning. We knew words mattered, but I’d not heard much talk at playtime. “If we’re not addressing personal, social, emotional development first, you’re not ready to learn,” said Prodger. She explained that before children could acquire the tools of speech and language, you had to ensure they felt a sense of “being and belonging”. Too frequently, she thought, our approaches to early learning skipped these steps. It sounded to me like a nice-to-have, not an essential, but research showed otherwise.
They take the kids out to the forest a few days a week, light fires, let them experiment with scissors and ride BMX bikes. If they want to be outside, they go outside. If they fancy returning to the snug, where the youngest infants roll around, that’s where they would go. The environment dictates the learning. The adults aim only to connect and share attention with the children. Reading and writing could wait. Nurseries ought to be as social as possible, and follow kids’ lead in their play. Before kids can get on with learning, we have to ensure they belong
“What the machine can’t do is be a partner,” Hirsh-Pasek told me. “It isn’t social. It’s interactive without being adaptive.”
He had discovered that human learning was communal and interactive. For a robot, the acquisition of language was abstract and formulaic. For us, it was embodied, emotive, subjective, quivering with life. The future of intelligence wouldn’t be found in our machines, but in the development of our own minds.
In 2003, the psychologist Patricia Kuhl experimented with teaching American infants Mandarin. Split into three groups (video, audio and flesh-and-blood teacher) only those with a human tutor learned anything at all. In 2010 a study of the wildly popular Baby Einstein vocabulary-building DVDs (Time called them “Crack for Babies”) revealed that infants who watched them “showed no greater understanding of words from the program than kids who never saw it”. Nor did babies learn words by eavesdropping on parental conversations or listening to In Our Time on Radio 4, however soothing the mellifluous tones of Melvyn Bragg. More than words, it took a human being for a baby to learn language. They could not learn from screens.
Erika Christakis, early-childhood expert and author of The Importance of Being Little, charts the slow descent in preschool learning from a multidimensional, ideas-based approach to a two-dimensional naming-and-labelling curriculum. Daphna Bassok at the University of Virginia asks if kindergarten is really the new first grade. The expectation that kindergarteners – aged five or six – can read is now commonplace. Yet this is counter to all the evidence. A Cambridge study comparing groups of children who started formal literacy lessons at five and seven found that starting two years earlier made no difference at all to a child’s reading ability aged 11, “but the children who started at five developed less-positive attitudes to reading, and showed poorer text comprehension than those who started later”.
These findings are clear: if you start on the decoding before you have an underlying understanding of story, experience, sensation and emotion, then you become a worse reader. And you like it less. Treat kids like robots during early learning and you put them off for life.
In her book, she suggested six Cs for modern learning: collaboration, communication, content, critical thinking, creative innovation and confidence. Truisms, I had thought, but unlike much education policy, drawn from scientific evidence. If I was to take away one thing, she said, it should be that “from the earliest ages, we learn from people”.
He had given up building robots that would compete with humans and instead turned his attention to the augmentation of human learning. What had changed his mind was the process of actually raising a child
