Spoken language has no clear pauses between , so how do babies know where one word ends and the begins? In a famous 1996 , Jenny Saffran and her colleagues played eight-month-old infants a of made-up syllables for just two minutes. Some syllables appeared together, others did not. Afterward, the babies listened to new combinations, showing that they noticed which syllables went . Babies, it seems, are natural .

Around the world, adults tend to speak to babies in a special way. They raise the pitch of their voices, exaggerate the rise and fall of their speech, speak more slowly, and repeat words often. Researchers call this infant-directed speech or, more informally, ‘parentese’.

For a long time, some people believed that talking to babies this way was silly or even harmful, because it did not model correct adult language. Research suggests the opposite. Babies prefer to listen to infant-directed speech, and its exaggerated sounds may make it easier for them to tell words and speech sounds apart.

The style appears in many languages and cultures, although it is not identical everywhere. A 2017 study found that mothers who spoke ten different languages, including Cantonese, Russian, and Hungarian, changed the quality of their voices in similar ways when addressing infants.

Researchers emphasize that what matters most is interaction. Speech from a television or a recording does far less for language learning than a real person responding to a baby’s sounds and gestures. Even very young infants seem to learn best through social engagement.