Kids Learn to Speak by Not Listening

Getting dressed in the dark is generally considered a bad idea. When presenting ourselves to the outside world, we like to have some visual feedback so we know what other people are seeing. Likewise, as young children learning to talk, we rely on auditory feedback--we need to hear ourselves speak. We continue to use this feedback, and adjust our talking accordingly, even as fully fluent adults. But at a certain stage of development, we may learn by not listening to ourselves at all.

Researchers led by Ewan MacDonald in Denmark tested Canadians of various ages on how they responded to auditory feedback while speaking. One group of subjects was adults; a second group consisted of kids around 4; and a third group was made up of toddlers around 2 years old. The study was simple: All the subjects had to do was say the word "bed" over and over. (The kids and toddlers were convinced to do this by playing a computer game. By saying "bed," they helped a distracted robot walk across a playground.)

Subjects wore headphones, and they heard normal feedback of their own words for the first 20 repetitions of "bed." But during the 30 repetitions that followed, the headphones began to lie to them. Subjects heard distorted feedback, so that what they pronounced as "bed" came back to them sounding more like "bad."

Adults and kids tried to compensate for this distorted feedback by changing the vowel they were saying.  Their "bed" became closer to "bid." But the toddlers didn't change what they were saying at all. They seemed to be speaking without listening to what they were saying. The robot was still crossing the playground, so visual feedback told them they were doing the right thing. Auditory feedback, on the other hand, was ignored.

You might expect toddlers to be listening a lot. But MacDonald speculates that when very young children are starting to talk, responding to the sound of their own voices might be counterproductive. Since toddlers' vocal cords are just getting accustomed to the sounds they need to make, listening to their own all-over-the-place babbling might not tell them much. Other kinds of feedback--say, visual responses from their parents--would give them a better idea of how they're doing.

Some support for this idea comes from language research in birds--not the kind that talk, but regular songbirds, which must learn complex vocal sequences in order to communicate and survive. MacDonald points to a study of brown-headed cowbirds, a small North American songbird, in which young males were raised by non-singing females. Without hearing songs to emulate, the young birds "nevertheless acquired mature, species-specific songs." Video recordings revealed that as the juvenile males sang experimentally, the adult females had given them visual feedback. This non-auditory response had been enough to help the birds learn. Like the cowbirds, human toddlers seem to rely on non-auditory cues for some part of their language learning.

Among the toddler group, hanging onto recruits for the duration of the experiment was a bit of a challenge. Out of 50 initial two-year-olds in the study, "Ten of the toddlers refused to talk and 13 refused to wear the headphones." Another six didn't speak consistently enough for the feedback system to work, and one toddler didn't manage to say "bed" at all.

This raises the possibility that the 20 toddlers who made it through the whole experiment were not representative of the group as a whole. They might have been the most proficient talkers, or the most mature. The authors argue that this shouldn't affect their results. Even if their subjects were the most mature toddlers, they say, those kids still didn't compensate for the vowels they were hearing, so the less mature talkers wouldn't have done any better.

But MacDonald doesn't address the possibility that these toddlers failed to compensate for altered vowels because they were more mature. Maybe talking without listening is a crucial developmental stage that must be reached, not just grown out of.

(Some adults, of course, seem to have never left the stage of talking without hearing themselves. Perhaps we should all stop nodding and smiling when they speak.)

MacDonald, E., Johnson, E., Forsythe, J., Plante, P., & Munhall, K. (2011). Children's Development of Self-Regulation in Speech Production Current Biology DOI: 10.1016/j.cub.2011.11.052