If adults speak so quickly (averaging 150 words per minute!) how do babies ever pick it up?

Far from being confounded by the slew of sounds they hear from birth (and even before), babies are slowly and carefully piecing together evidence of how speaking works, what it's for, how it feels physically and emotionally, what all these sounds mean, and how to distinguish between them. Perceiving and distinguishing meaningful sounds is the beginning of learning to speak.

Different languages divide up the continuum of sounds in different places. If you listen to a /r/ sound and a /l/ sound, and you're used to hearing English, they sound distinguishably different. Rap and lap mean totally different things, right? This is called categorical perception: we group similar things together but we also filter out any minor differences in how different people say a sound. So babies hear /p/ in 'pat' from different adults and perceive one sound for /p/ not a new sound for each speaker (boy, that would be confusing!) Categorical perception isn't unique to humans: researchers have trained chinchillas to do it too (Kuhl & Miller, 1975). Human babies do it naturally. No training needed.

Different languages group sounds together in different ways. In Japanese, the sounds /r/ and /l/ are not perceived as being different sounds (called phonemes) but as variations of the *same* phoneme. These are called allophones or free variants of the same sound. This is why it's harder for Japanese adults to hear the difference between /r/ and /l/.

Sounds can also differ slightly depending on where they are in a word, like the /p/ in map or petal. Languages carve up the continuum of sound in different places. Babies take all this language variation in their stride.

Babies are very 'open' to learning where these boundaries may or may not be in the languages they hear. Zulu click sounds carry meaning for speakers of Zulu, but are not meaningful speech sounds for English adults who do not speak Zulu. They might sound no more meaningful than clicking your fingers, and all sound the same at first, unless you practise hearing the differences. But English babies of 6–14 months old can perceive the contrasts between Zulu click sounds, according to research by Best and colleagues in 1988. Adults could also perceive the difference: the sensitivity to subtle differences was still there. Maybe adults just need to have confidence in their own listening abilities – something I think we need to work on!

It's really a matter of perspective. Learning about how babies carve up the speech sounds they hear into meaningful chunks, slowly over a year or so of listening, is really incredible. Keep an eye out for more posts about what happens next in this seemingly miraculous process!