Sunday, April 27, 2008

Slow it down, crank it up

I have been re-reading Philippe Ariès' Centuries of Childhood, the classic account of the development of concepts of childhood and family in medieval and early modern France. One source Ariès makes use of is the medieval encyclopaedia Le Grand Propriétaire de toutes choses, compiled in Latin in the thirteenth century. In the 1556 French edition of the work, Ariès finds this description of a nanny:
"She rejoices when the child is happy, and feels sorry for the child when he is ill; she picks him up when he falls, she binds him when he tosses about, and she washes and cleans him when he is dirty."
What's more, the medieval nanny has a particular way with language: 
"She pronounces the words as if she had a stammer, to teach him to talk better and more rapidly..."
In other words, she is doing this:

This mention in the Grand Propriétaire must be one of the earliest descriptions of motherese. In the Confessions, St Augustine famously introspected and retrospected on his own learning of language as an infant, part of which involved understanding the intentions of those around him. He didn't, though, suggest that any of the adults around him deliberately set out to make his language acquisition any easier. Do examples of this kind of baby-talk appear any earlier, such as in the literatures of ancient Greece and Rome? Please comment if you know of any precursors to this depiction mentioned by Ariès.

It used to be thought that motherese was a special language, with its own grammar and structural rules. The 'stammer' mentioned by the medieval compilers of the Grand Propriétaire refers to the tendency of speakers of motherese to slow their speech down, to use intensive repetitions, and to emphasise changes in pitch in that classic 'sing-song' way. Today, motherese is recognised as a particularly effectively way of communicating with infants at the same time as teaching them about that medium of communication. In Chapter 7 of The Baby in the Mirror, I describe how babies use statistical regularities in the language they hear spoken around them to build up knowledge of the phonology of their native language, to understand word segmentation, and so on. By emphasising pitch changes and prosodic patterns, caregivers facilitate this process. It might make non-parents cringe, but it is doing an important job. 

And it doesn't just work in the realm of language, either. Researchers at Villanova University have shown that parents use a form of motherese in their object-centred interactions with babies. Understanding human actions poses the same kinds of challenges as understanding language: you have to break a fluent sequence of behaviours down into intentions and action boundaries, and 'parse' it as you would parse a sentence. In the first study of its kind, the researchers asked adults to demonstrate the functions of unfamiliar toys to either babies or adults. They found that, when demonstrating to infants, the participants would modify their actions so as to emphasise relevant parts of the action. Such demonstrations were more interactive and enthusiastic, simple and repetitive, and performed in closer proximity to the individual they were intended to benefit. 

We don't yet know whether this 'motionese', as the researchers dub it, has any causal role in helping babies to learn about actions. Researchers would have to show that differences between caregivers in their willingness to use motionese translated into individual differences in babies' understanding of action sequences. But it seems highly likely that these kinds of modification of behaviour will provide useful input for developing brains. Next time you see someone overdoing it for the benefit of a baby, don't be too hard on them. They will be performing one of the most useful functions a caregiver can fulfil: slowing and simplifying the flow of action for a person who has much to learn. The baby will probably be loving it, too.