The Articulate Mammal (29 page)

I got it;

my disability;

not never to walk from it.

It shares my space,

breathes the same air.

I cannot have a day off.

Or, as she said in another poem:

I lost the me

It got under everything

that was not poems.

Christopher and Kate show that language can not only be spared, but even enriched, when other cognitive abilities are impaired. So far, then, we have argued that language seems to be a special skill. In occasional cases, it can be partially separated from general cognitive ability.

But normal children do not just acquire language by lucky chance. They exploit the helpfulness of those around, and (mostly) pay attention to what they say.

CAREGIVER LANGUAGE

NO, YOU SHOULDN’T MAKE A HOLE IN THE BOTTOM.

NO, LET DADDY DO IT FIRST.

NO, I DON’T THINK YOU’LL BE ABLE TO CUT STRAIGHT.

NO, DON’T CUT TOO MUCH ON THE FRONT.

NO, I DON’T THINK YOU SHOULD PUT MORE HOLES IN IT.

NO, DON’T CUT TWO HOLES.

These six examples of NO placed in front of the sentence occurred in a 15-minute recording session in which 2-year-old Nicholas was ‘helping’ his father to carve a pumpkin (De Villiers and De Villiers 1979). This construction seemed to be a favourite one for Nicholas’s father. Not surprisingly, the majority of Nicholas’s early negatives (76 per cent) involved opposition to some suggestion proposed by his parents, and his means of expressing this was by placing NO in front of the sentence, as in NO DADDY DRESS ME.

Faced with such examples, a number of people have suggested that
motherese, caretaker language
, or
caregiver language
(speech addressed to children) can solve the mystery of how children acquire language so efficiently. Children, according to this view, absorb and copy the speech they hear around them. They learn so fast, it is claimed, because speech addressed to children is rather different from that to adults (
Chapter 4
), and so grabs their attention. In many communities, this special way of talking to children begins as soon as the baby is born:

Wha’s a matter, Bobby, yo’ widdle tum-tum all empty? Here you are, a growin’ boy, and dese folks won’t feed you. You tell ‘em, they can’t just let you cry, not while Aunt Sue is ‘round … You’re a-gonna be a big boy, just like your daddy. Mamma gonna hafta get some new rompers soon … Okay, okay, look, look, there’s mamma, she’s comin’, she gonna get dat bottle right now and get it ready for you. It’s a hungry boy, it is.

(Heath 1983: 118)

This stream of speech was addressed to Bobby by a helpful neighbour when he was only a month or so old. It contains many of the characteristics found repeatedly in child-directed speech (Ferguson 1978). It tends to be slower, spoken with higher pitch, and with exaggerated intonation contours. The utterances are shorter, with the average length being approximately one-third of that found in speech addressed to adults (Newport, Gleitman and Gleitman 1977). The sentences are well-formed, simple in structure and repetitious, in that the same lexical items recur, though in slightly different combinations. Special ‘baby’ words are sometimes used, such as DOGGIE, BIRDIE, GEE-GEE, CHUFF-CHUFF, TUM-TUM. The topic is usually related to the ‘here and now’ – things that are present both in place and time.

Yet the link between caregiver speech and child language is not always straightforward. And ‘repairs’ – cases in which caregivers try to ‘mend’ a communication which has been ignored or misinterpreted – turn out to be too infrequent to be of consistent use. Meanwhile, reformulations such as:

OPEN YOUR MOUTH. OPEN IT.

SPIT OUT THE SNAIL, SPIT THE SNAIL OUT. SPIT IT OUT.

GIVE MUMMY THE SNAIL. GIVE THE SNAIL TO MUMMY.

account for only around 4 per cent of maternal speech, it has been claimed (Shatz 1982).

So is Chomsky right in his belief (outlined in
Chapter 5
) that what children hear is ‘fairly degenerate in quality’ (1965: 31)? Or has he never listened to parent–child conversation? Let us consider the matter more carefully.

For a start, it seems reasonable to expect that words and constructions which occur frequently in adult speech will be produced early by children. And this certainly seems to be borne out in some studies. For example, in the development of verbal auxiliaries (words such as
can, will, might, have
, etc.) the order of acquisition roughly follows the frequency of these words in adult speech (Wells 1979).

Statistically, therefore, there is a link between items produced frequently by parents, and those acquired early by the child. But the problem with statistical correlations is that they do not hold for every construction, nor for every child. Furthermore, correlations which are valid for groups of people can sometimes disappear when each individual child and its parents are examined separately (Wells 1979, 1986). We must conclude, therefore, that overall
frequency of use is only a rough guide to the order of acquisition, and is by no means a definitive map.

Since simple frequency counts have not proved entirely helpful, some researchers have suggested that motherese directs child language in a more subtle way. They have proposed that parents have an inbuilt sensitivity to their children. According to this view, parents gradually increase the complexity of their speech as the child becomes ready for each new stage. This has sometimes been called the ‘fine-tuning’ hypothesis (Cross 1977), in the sense that parents subconsciously attune their output to their child’s needs. And a few people have claimed that, far from children possessing an innate language learning device, mothers possess an innate language teaching device! Those who support this viewpoint assume that there will be a close correlation between the structure of the mother’s speech and that of the child at every stage of development. Is this true?

Research confirms that parents attune their speech to their children’s needs, but suggests they attune them to a child’s interests, more than to his or her language structure. That is, parents talk about topics which are relevant to the child such as picking up blocks or drinking juice, but show no evidence that they are grading their syntax, or introducing constructions one at a time, as one might expect if they were subconsciously guiding their children from one stage of language to another. There is no sign of a step-by-step programme, except in the broad general sense that as the child gets older, the parents’ speech tends to become less repetitious, with longer sentences and more complex subject matter. Moreover, researchers who examined the speech of fifteen mothers interacting with their young daughters, concluded that if one was designing a curriculum for language teaching, motherese was highly unsuitable! (Newport
et al.
1977). In a good language teaching programme, you would expect teachers to introduce constructions one at a time and to concentrate first on simple active declarative sentences (TOBY WANTS A BATH, or MARION IS EATING A BUN), then move on to constructions in which words are omitted or the order shifted round as in imperatives (TURN OFF THE TAP! COME HOME!) or questions (WHAT IS TONY EATING? WHY ARE YOU CRYING?). Instead, they found mothers did the reverse. That is, they used all these constructions jumbled together with more questions and imperatives (62 per cent) than declaratives (30 per cent)! Oddly enough, there were more declaratives in the
second
session, six months later, than in the first, and even more in the speech addressed to other adults! These researchers, therefore, assert that ‘Motherese is not a syntax-teaching language.’

Children, therefore, have an inbuilt filter which allows them to choose what they pay attention to: ‘The child is selective in WHAT he uses from the
environment provided; he is selective about WHEN in the course of acquisition he chooses to use it’ (Gleitman
et al.
1984). Child ‘uptake’ is not matched in any straightforward way with adult input.

A further piece of evidence that uptake matters more than input, is the existence of communities where parents do not modify their speech when talking to infants. In ‘Trackton’, a working-class black community in the southeast of the USA, adults do not regard babies as suitable partners for regular conversation (Heath 1983). They rarely address speech specifically to very young children. Moreover, Trackton inhabitants find it odd when they hear white people gurgling over infants: ‘White folks uh hear dey kids say sump’n, dey say it back to ‘em’ (Heath 1983: 84). Trackton children are an integral part of family life, so they hear plenty of speech around them. Somehow, they acquire language as efficiently as anyone else.

Adults can, however, help their children by talking about things that interest them, and engaging in joint enterprises (Wells 1979): ‘Now, then, Shirley, are you going to help mummy peel the potatoes? Can you get me six out of that basket?’ The tendency of girls to be mildly ahead of boys in their language may be due to the different treatment meted out by parents. Girls are often kept in to help with the chores in many families, but boys are sent out to play games. Mothers, rather than footballs, aid progress in language. And sensitive fathers can help their sons too.

Some researchers have queried why the role of caregivers arouses such controversy among those who study it. One finding is that the various types of interaction have been classified in overlapping and confusing ways. Researchers may, as a preliminary step, need to distinguish negative
feedback
from negative
evidence
(Saxton 2000).

Negative feedback is a signal to the child that something is wrong with their utterance, though exactly what is unspecified:

This query prompted Alex to reassess what he had said, and amend this too: ‘The pirate hit him on the head’ (Saxton 2000: 228) – though Alex must already have known the correct form, because no extra information was provided by his father. Negative evidence, on the other hand, provides the child with the information needed to put the utterance right:

Where does all this lead us? We now realize that caregiver speech is an important factor which must be taken into consideration when we study child language, and that it is often considerably more coherent and more useful than Chomsky suggests. Yet as Cromer (1981: 65) notes:

To show that the input signal to many children is far clearer than had been assumed in no way explains how the grammatical structures that the child uses are developed.

Or, in the words of the nineteenth-century German philosopher-linguist Wilhelm von Humboldt: ‘Language cannot really be taught … One can only offer the thread along which language develops on its own’ (quoted in Slobin 1975: 283).

A LINGUISTIC PROCESS PEGGY

Process Peggy, then, is not a general problem-solver. Neither everyday needs nor general intelligence nor caregiver speech can fully account for her special language abilities – though all these factors are important if she is to develop normally. She must be innately programmed to tackle language. In this section, we shall discuss how she might set about her linguistic puzzle-solving.

In the beginning, she possibly uses her general intelligence to get going (
Chapter 6
). She may behave like a computer, which often needs a fairly general program to start up before it can use a more specific one. Computer operators talk about ‘booting up’ or ‘boot-strapping’ a computer – giving it some preliminary commands, which will then allow it to cope with more detailed programs. So some linguists talk about a ‘bootstrapping’ approach to language (Pinker 1984, 1987).

Linguistic bootstrapping might work as follows. Children learn words such as DOGGY, KITTY, BITE, DRINK, BALL, MILK, which correlate well with actors, actions and objects. They therefore build these up in various semantic relationships (
Chapter 6
):

KITTY DRINK (ACTOR + ACTION).

DRINK MILK (ACTION + OBJECT).

And they may combine these into longer sequences:

KITTY DRINK MILK (ACTOR + ACTION + OBJECT).

Up to this point, general intelligence may be at work, rather than linguistic ability. The basic scaffolding relies on meaning.

Then they switch over to syntax. Exactly how they do this is disputed. According to one suggestion, syntax begins when children discover some discrepancy in their semantic scaffolding. They may discover that words describing actions such as BITE or DRINK can sometimes be replaced by words such as WANT, GOT, LIKE, which do not involve any kind of deed:

KITTY WANT MILK.

KITTY LIKE MILK.