How Do We Know What Babies Know?

Babies, it seems, do something new every single day. Before they are even able to control their own body movements, they are engaged in a furious observation of the world around them, learning from just about everything they see, hear, touch, taste, and smell. Although there is a wide range of “normal,” of course, lots of studies have determined which concepts babies understand at certain ages and stages. For example, there is evidence that babies recognize their mother’s face in as little as three days after birth and that they have a keenly developed number sense by six months of age.

_-0-_ via Flickr

But how can scientists determine what babies know? Asking a newborn whether the woman holding him looks like his mother isn't likely to elicit a useful response. One way to get a read on babies’ thought patterns is by performing a brain scan. But this technique isn't ideal. For example, a functional magnetic resonance image (fMRI) provides great images but is very expensive, and electroencephalography (EEG) is cheaper but doesn't yield information about impulses deep within the brain.

Instead of measuring brain activity, researchers have learned that babies’ behavior in response to stimuli can reveal a surprising amount of information about what goes on inside their heads. Most babies respond in predictable ways to novelty and many experiments take advantage of this tendency. Careful observation of babies has led to some useful, measurable methods that help researchers as they work to probe the minds of the very youngest people around.

Linguists, for example, are interested in the way very young babies hear sounds. Speakers of certain languages often cannot hear the difference between sounds in a foreign language; for example, Japanese and Chinese speakers struggle to differentiate between the /r/ and /l/ sounds, and /v/ and /w/ sound the same to speakers of Hindi and Thai. Are babies hardwired to speak a particular language from the moment they’re born, or is this selective deafness a learned trait?*

To determine what babies hear, scientists use the concept of novelty. Babies are provided with a high amplitude sucking device to measure their responses, then are exposed to sounds. To the baby, a high amplitude sucking device feels like a pacifier, but in fact it is connected to a system that measures the rate at which the baby sucks it. During the experiment, the baby listens to a recording of one of the target sounds over and over again. For a baby born in an Arabic-speaking environment, for example, researchers might choose to play “pah, pah, pah...” When the sound begins, the baby will begin to suck the pacifier at a faster rate, but as it grows accustomed to the sound, it will demonstrate its boredom by sucking more slowly. Then, suddenly, the recording will change; the baby will begin to hear “bah, bah, bah...” Adults and even young children who speak Arabic have great difficulty hearing the difference between /p/ and /b/, and most don’t notice when “pah” switches to “bah.” But most babies begin sucking much faster the instant the sound changes; their curiosity is aroused by the difference, and they become attentive and interested. To their older counterparts, the stimulus appears unchanged, but babies demonstrate a much keener sense of sound discrimination.

Another way to measure a baby’s perception is to record the amount of time she spends looking at something. Just as babies suck faster when they hear something new, they tend to look longer at things that are different from what they know or that violate their expectations. This concept, known as “preferential looking,” was first developed in the 1960s, and scientists still use it today. Researchers at Johns Hopkins University, for example, demonstrated that babies understand the limitations of the physical world by showing them a series events and recording their looking times. Some of the events were deemed “real,” in that they were possible, but some “magic” events were physically impossible. In one “real” video, a ball rolled up to a wall and bounced off. Babies weren't too captivated by that, but they couldn't take their eyes off a “magic” video of a ball rolling into, then through, a solid wall. This fascination indicates that babies as young as two and a half months had gained a great deal of knowledge about the physical properties of the world.

Babies may spend a lot of time gazing around them, reaching for objects, listening intently, knocking things over, and banging on surfaces. But the serious learning going on beneath the surface is anything but child’s play.


*Interestingly, the reason older children and adults can’t hear the difference between similar sounds in other languages has nothing to do with their hearing. The answer is found, instead, in the brain. Babies are born with more than 80 billion neurons (brain cells) and synapses (connections between brain cells) in their brains – more than are found in any adult. This means that they are prepared to detect all kinds of stimuli in the world. The problem, however, is that having lots of extraneous neurons and synapses means that information and impulses don’t travel very quickly. Imagine searching through a suitcase for a particular item; if the suitcase is filled with things, it takes a long time to find what one wants. Similarly, the multitude of structures in the infant brain can make it work more slowly. Just as taking out half of what’s in the suitcase can cut down on search time, babies’ brains reduce the synapses in the name of efficiency. As the baby observes and interacts with its environment, the cellular connections that aren’t needed get the boot. This regulatory process is known as pruning. So after spending a year or so in a Chinese-speaking environment, where the /r/ and /l/ sounds don’t contribute to meaning, the brain of an infant there would determine that being able to detect the difference between the two sounds is unimportant. That synapse is pruned, making way for more efficient synaptic connections that the baby has noticed are relevant. Pruning is thought to result in learning, as the brain customizes itself to perform optimally according to observed environmental factors.