Motherese can be heard as a mother talks to her baby about toys.
“Motherese” is the simplified, slowed-down, high-pitched speech that people use when talking to infants and toddlers. Even slightly older children use it when speaking to infants, and of course mothers use it most of all. Motherese syntax is simplified too. Dada go bye-bye is recognizable motherese. But does motherese have anything to do with a child learning a language, or with the evolutionary origins of speech? In her new book, Finding our Tongues Dean Falk makes the case for its relevance. She rejects the argument that it is merely a device for strengthening the mother-infant bond and, despite the fact that previously I have tended to dismiss the motherese argument, I think her case is a good one.
The chief argument against the importance of motherese is that it is not universal to all cultures and all mothers. Falk’s reply is twofold: first, it is pretty much universal, and, second, in the cases where it is absent children are slower to learn to speak. The three most cited examples of non-motherese-using cultures are in Western Samoa, New Guinea, and among black, working-class Americans, so Falk investigated and found:
- Western Samoa: The Upolu people of Western Samoa live in a hierarchical society in which language use reflects social ranking. Although it is widely asserted that the Upolu do not use motherese, Falk says the data supports the opposite conclusion. “Language addressed to the young infant tends to be in the form of songs or rhythmic vocalizations in a soft, high pitch … [but this ends dramatically] once a child is able to locomote” [p. 92]
- New Guinea: The Kaluli tribe has an egalitarian system in which babies are placed to view the whole gathering rather than focus on the mother. Speech tends to be speaker-to-listening-group rather than to a single listener. Particularly striking is the way the mother speaks for the infant when it is too young to speak for itself. These exchanges take place in motherese with the mother speaking in a high-pitched, simplified style and older children also speaking motherese. Falk’s conclusion is that the Kaluli do have motherese, although much adapted to their cultural conditions.
- Working class African-Americans: Again Falk counters data with more recent data. In this case, she responds to data about African-Americans in South Carolina with data from Columbus, Ohio. It defeats generalizing from South Carolina to the whole, but still doesn’t settle the issue of what is going on in South Carolina.
Falk does a pretty good job of arguing that closer looks tend to support the presence of motherese, with perhaps some culturally-imposed oddities. More telling was her general complaint that, “When speculating about motherese, many anthropological linguists focus on the conscious teaching of language to toddlers, a very late state in language acquisition” . That’s a stronger point in defense of her position: observers have been looking in the wrong area. The Upolu people, for example, quit using motherese just before linguists start looking for it.
Motherese was reported and first studied in the 1970s, when linguists were most deeply enthralled with the generative theories that language was characterized by the syntactical rules that define sentences. It quickly became apparent that motherese has little to contribute to proper syntax and the subject was dismissed as having something to do with mother-child bonding. But speech requires other talents, notably the ability to perceive the sounds of speech, to attend to the differences that matter and ignore those that do not, to notice when words are replaced by other words, and to recognize the variations that indicate things like verb conjugation, pluralization, etc.
Anyone who has tried to learn a foreign language knows how difficult it is to hear these details as they whizz by the ear. We usually say that the knowledge of one language interferes with the perception of the second, but that can’t be the cause of any problems infants face. Even so, infants can become part of the group only by learning to perceive speech sounds the way their elders do.
Most perception thresholds appear to be innate. We see color boundaries, we do not learn them. It is the same for pitch differences, but word boundaries have to be learned and they work very differently from culture to culture. Experiments have established that very young infants can distinguish between all the vowels of the world’s languages, but they lose that ability as they begin to master a particular language.
Falk argues that motherese is crucial during the first year at providing infants with the experience necessary to perceive a language. It exaggerates certain syllables within words and certain words within sentences, helping newborns “discover more easily how to divide speech into words and clauses long before they learn their meanings.”  What’s more,
We haven’t paid close enough attention to language perception and we have no good measurements of how well people perceive speech. We just assume kids in school can hear the words spoken by teachers and classmates. We don’t really know if they can, unless they are extreme cases, such as autistic students who do not perceive language as language. Attention to prosody and language’s musical qualities is the greatest strength Falk brings to her book. I have not paid much attention to this issue before, but it is obvious now that part of the origins of speech included the rise of the ability to perceive the words in the stream of speech. (Falk also reports a kind of motherese in sign language that helps children learn to perceive the word signs in the stream of signing.)
Why did the human lineage first begin using motherese? Falk says she ‘pondered this problem for some years before formulating the ‘putting the baby down’ hypothesis”  which I discussed in my first post on Falk’s book (see: Where Shall I Put the Baby) and which I find quite unpersuasive. I also find it unnecessary. Falk says she believes motherese emerged before either protolanguage or true language, and that makes sense to me. The first speakers had to be able to distinguish words, so they had the same challenge that infants today have in learning to recognize the speech sounds of their group. It is possible that the first speakers had the capacity to learn to recognize words when spoken at a modern pace, but there is no reason to assume such power. So the simplified sentences and exaggerated words may have characterized the first speech of adults as well as infants. Maybe two million years ago adults said things like Dada go hunt-hunt to other adults.