As if being born weren't hard enough, as if all the strange new sights and sensations weren't confusing enough, infants are hit with a particularly difficult problem. They need to learn language.

Breaking Sound Into Discrete Chunks

This task is especially hard because we don't speak slowly, and clearly. Instead, our speech generally leaves no obvious pauses between words, creating a sort of double bind for infants. Before they can learn what the words of a language mean, they need to figure out what the individual words are. That is, they need to break a stream of sound into discrete chunks, and they need to do this without knowing what these chunks are ahead of time.

Identifying Boundaries

How is it done? One likely theory, proposed by Richard Aslin of the University of Rochester, suggests that babies get used to the types of syllables that are likely to follow each other within a word. When they hear a pair of syllables that don't seem to belong together, they soon realize that this might indicate a boundary between words.

Aslin's example is the phrase "pretty baby." After the first syllable, "prih," we are more likely to hear a syllable like "tee" in an English word than, for example, "pop" or "bay." Likewise, "bay" is likely to be followed by "bee." Between "pretty baby," however, we get the combination "tee-bay," which is not at all likely in an English word. By recognizing which syllable combinations are common, and which are uncommon, babies can begin to recognize the breaks between words.