How Babies Work

How Babies Learn to Be Human

A woman asks her child to pay for vegetables at a food market in Shanghai, China on March 9, 2013.

A woman asks her child to pay for vegetables at a food market in Shanghai.

Photo by Aly Song/Reuters

It is hard to imagine being a baby, and not just because it is hard to imagine not knowing what you already know. That’s the easy part. It is far harder to imagine not knowing what you do not know you already know.

A sentence that Rumsfeldian cries out for a hypothetical: Let’s say you are in line at Starbucks and the barista puts a cup of coffee on the counter. As an adult, you understand that action as: That person put a cup of coffee on the counter for a customer. As a baby, or at least a newborn, you perceive it as: That person extended his arm through space while holding an object and then released the object.

In short, there was a moment in your life when you did not understand the idea of intentions. When a person acts, she has a goal in mind—she intends to do something. This is the insight that undergirds your understanding of almost everyone you see around you; it is the insight that makes sense of the social world. It’s how you know someone wants to shake your hand when he extends his toward you. But it turns out babies aren’t born with this insight. They have to acquire it—and until they do, the baristas at Starbucks make absolutely no sense whatsoever.

In recent years, developmental psychologists have begun to figure out how and when infants connect intention to action. (And other insights that are equally crucial: that what someone is paying attention to is a clue to their intentions; and that intentions belong to individuals—that not everyone is reaching out their arm because they are putting down some coffee.) The answers are fascinating: Infants learn about the social world at least in part by extrapolating from their own actions. In other words, if a baby was inadvisably hired at Starbucks and had the experience of putting a cup of coffee on the counter for a customer, she would be far more likely to understand what the other baristas were doing.

A lot of the most interesting work in this field has come out of the lab of Amanda Woodward, a psychologist at the University of Chicago who came to the subject after studying language development in graduate school. Her experience led her to wonder less about language and more about what happens before language. Her subjects already possessed an enormous amount of social intelligence. In learning words from her, they were clearly trying to divine her intentions: What was this strange graduate student trying to do? So Woodward decided to study how babies get to that point. What happens that enables that year-old infant to understand the idea of intentions?

After a mere half-year in the world, babies have figured out some version of how it works. In a visual habituation experiment—a key method in infancy research, which exploits the fact that infants look longer at events they perceive as novel—the 6-month-olds looked longer when a person reached for a different object but not when they reached in a different direction. They understood that the object, not the reaching itself, was what mattered most. But when the babies watched a mechanical claw reach for a different object, they failed to look longer. They expected humans to have goals; they had no such expectations for claws.  

How do they come to understand this? They mine it from their own experiences down there on the floor. The best example of this sort of learning comes from an experiment involving “sticky mittens,” Velcro-covered mittens that allow infants to pick up toys that are also Velcro-covered. (The fact that developmental psychologists sit around and think up experiments involving things called “sticky mittens” should really inspire a lot of people to go to graduate school.)

At three months, babies aren’t very good at reaching and they aren’t very good at understanding what other people are doing by reaching. But after they wear the “sticky mittens” and have the experience of picking up toys, they suddenly understand the point of reaching. In another visual habituation experiment, they now notice when the experimenter reaches for a new object—that is, when the experimenter has a new goal in mind. (The babies who watched the experiment first and then played with the mittens didn’t notice.) The longer a baby played with the mittens, the stronger the effect was. Just a few minutes of active experience altered how that infant comprehended the world. This appears to be true for more sophisticated actions, too. At 10 months, babies are good at reaching but they’re not very good at higher-level, means-ends actions. However, when infants this age learn to use a tool—in this experiment, using a cane to get a toy—they then understand what adults are intending to do when they use the tool. They focus on the ends (the toy), not simply the means (the cane).

It would make sense that babies, after discovering that people have these things called goals, start seeing goals everywhere they look. As Woodward says, “You might think that babies start with a kind of crude expectation—that anything a person does is probably goal-directed.” But, she continues, “that’s not the way it is.” When 7-month-old babies see an experimenter reach toward one of two objects, and then are given the opportunity to reach toward those same objects, they unsurprisingly choose the same object as the experimenter. But when the experimenter reaches more ambiguously—when she just touches the toy with the back of her hand, for example—the infants choose randomly between the toys. They don’t interpret her movement as having a goal. It’s a remarkably fine-grained understanding of actions.

And it is achieved in a remarkably short period of time. Woodward’s work highlights just how amazing infants are at learning: Their achievements never seem less than wondrous. And Woodward herself was expecting less. She’d done postdoctoral work with Elizabeth Spelke, who revolutionized infancy research with her theory of infant core knowledge—the idea that we are born with innate capacities that we get for free. “I assumed that this piece of social knowledge would look just like the rest of infant core knowledge, that it wouldn’t depend on experience,” Woodward says. “And so I was initially frustrated at why only older babies would pass my measures.” It turned out that before they could pass, they’d had to study up: They had spent their brief time in the world cramming.

The idea that babies can be smart and still need to study shouldn’t be surprising. But it is. “Bizarrely, in the history of the field, there’s been this conjunction of If learning is important, babies must not be smart, and If babies are smart, learning must not be important,” Woodward says. “It’s just a logical fallacy.” Babies have to be smart in order to learn what they do.

Innateness now dominates the popular understanding of infancy, and clearly infants have extraordinary inborn knowledge. As innumerable studies have now proved, babies can think in very sophisticated, abstract ways. But we don’t get everything for free. As a parent, I find this warp-speed learning even more marvelous than innateness. It’s wonderful to watch my infant son and know that he is multitasking: He is drooling and deciphering the foundations of human social life. He is learning things that I didn’t even know I knew. And this learning isn’t passive; it is very much active. You can watch as babies, almost literally, piece together the social world with their hands.


Nicholas Day’s book on the science and history of infancy, Baby Meets World, was just published. His website is He is @nicksday on Twitter.