One of the tasks facing a child learning his or her language is to figure out the sound system. This involves learning how to distinguish all the linguistically important differences, and also how to produce them. It's rather easier to record what small children say than to determine what they understand, so most systematic research has examined production. At the same time, it is widely believed that children's phonological perception runs ahead of their productive abilities, and this mismatch between perception and production will take on considerable significance as our discussion proceeds. Because most of the relevant research has been conducted on English-speaking children, we shall restrict ourselves to the acquisition of English.
It is remarkable that children seem to be innately disposed to perceive the sounds of language. In an ingenious series of experiments, Peter Eimas and his colleagues have shown that very young babies can hear the sorts of distinctions that are often used in languages and to which we have given some attention in the previous section. The techniques revolve around one idea: a baby quickly gets bored unless something different happens in its environment. Experimenters therefore play a series of identical sounds to a baby, say the syllable [pa]. At first the baby is interested and turns its head to the sound. As the sounds are repeated, it loses interest and stops turning its head.