The Articulate Mammal (23 page)

He also claimed that the same child understood the notion of location because she pointed to the empty top of the refrigerator, where bananas were normally kept, and said NANA. He concluded that ‘there is a constant emergence of new grammatical relations, even though no utterance is ever longer than one word’ (McNeill 1970: 23). So, McNeill might perhaps suggest that BA, when applied to kitchen taps and milk, showed an understanding of location: ‘There are taps like this on the bath tub’, ‘There is liquid like this in the bath.’

McNeill claimed that children understand a wide variety of grammatical relationships and that one-word utterances are sentences in embryo. This seems over-imaginative to many researchers. However, the idea that single-word
utterances may be more than mere labels has also been examined by Lois Bloom, a researcher at Columbia University, who put forward a fourth possibility (Bloom 1973).

After a careful analysis of the single words spoken by her daughter, Allison, she suggested that there is no simple answer to the problem of interpretation because the meaning of a one-word utterance varies according to the age of the child. For example, when Allison said MUMMY at the age of 16 months, she seemed to mean, simply, ‘That’s Mummy’. But at the age of 19 months she appeared to be trying to express some kind of interaction between Mummy and the surrounding environment, as when she pointed to her mother’s cup and said, MUMMY.

However, Bloom was unable to tell exactly what kind of interaction was intended. Did Allison mean ‘That’s Mummy’s cup’, or was she saying ‘Mummy’s drinking from a cup too’? Because of this intrinsic ambiguity, Bloom was cautious about assigning specific meanings to BA-type words which relate either to objects, or to interaction between objects. She was more optimistic about the interpretation of words such as NO, MORE and A’GONE in which ‘conceptual notions are so conveniently tied to the actual words in the child’s speech’ (Bloom 1973: 140). For example, Allison showed by her use of the words NO and A’GONE that she could cope with the notion of non-existence. Bloom, therefore, concluded (perhaps not surprisingly) that single words are grammatically fairly uninteresting. Their importance lies in the light they throw on a child’s conceptual representation of experience.

Other researchers have tried to analyse what the child is trying to
do
with one-word utterances (e.g. Wells 1974; Halliday 1975; Griffiths 1986). If a child says GA, is she simply naming an object such as a cat? Is she asking for the cat? Or is she trying to control the actions of her parents by telling them to let the cat in? All of these are possible. The probable ‘translation’ may even depend on the temperament of the child. Some children simply enjoy naming things, others prefer to use words to get the attention of the adults around them.

An extra reason for caution is that some youngsters may not even realize that they are ‘naming’ things when they first utter words (McShane 1979, 1980). They may simply be taking part in a ritual game. Many middle-class parents sit down with their children and leaf through picture books, naming the objects which appear on each page, such as ‘apple’, ‘ball’, ‘cat’, ‘duck’, and so on. The child may shriek BA delightedly when she reaches the page with the round blue blob in the middle, but may not for some weeks realize that this sequence of sounds is actually the ‘name’ of a certain type of round object, a ball. As McShane put it: ‘The child first learns the words and later learns that these words are names’ (1979: 890). The sudden realization that things have names appears to lead to a surge of ‘labelling’ everyday objects
such as CAR, MILK, BALL, APPLE, followed by a surge of ‘describing’, with the use of words such as BLUE, GONE, BROKE, HIT. This in turn, he suggested, leads to the beginning of structured speech. His suggestion is supported by others: ‘This burgeoning store of comprehended words triggers or reinforces the activation of analytical mechanisms’ (Locke 1997: 276).

But the situation is by no means clear cut. Some children go through this sudden ‘naming insight’ stage, others seem to know that things have names before they start to utter any words (Harris
et al.
1988). And occasionally, the ‘single word’ stage may even be missed out. There are reports of a workingclass black community in Pennsylvania where it is considered odd to talk to babies, and parents make no attempt to interpret children’s early babbles as labels. These children often begin to communicate by picking up whole phrases, which they use with a wide range of intonations and meanings. One toddler, Teegie, used ‘You shut up’ to mean ‘No’, ‘Leave me alone’, ‘Give me that’, and ‘Take it, I don’t want it’ (Heath 1983). But these children learned language perfectly well via this route. A young child is ‘faced with having to discover what talking is all about’ (Griffiths 1986: 281), and there seems to be no one way in which this realization comes about.

This brief excursus has by no means exhausted the views on one-word utterances found in the literature. It does, however, illustrate one important point: when the data are so confusing, it is no wonder that differences of opinion abound in child language studies. All researchers, to some extent, see what they want to see. This accounts for the extraordinarily diverse viewpoints which arise over apparently simple issues.

Having pointed out the type of problem that is likely to arise, we must now return to the main topic of this chapter, which is the following: how do children get started on learning to speak? We shall consider this question by looking first at children’s two-word utterances. We shall then go on to examine how children acquire more complicated aspects of language such as word endings and negation.

TWO-BY-TWO

There are basically two ways of analysing two-word utterances. We may choose either the ‘Let’s pretend they’re talking Martian’ technique or the ‘Let’s guess what they’re trying to say’ method. In the first, linguists approach the child’s speech as if it were an unknown exotic language. Having freed their minds of preconceived notions connected with their knowledge of English, they write a grammar based entirely on the word patterns they discern in the child’s speech. In the second method, linguists try to provide an interpretation of what the child is saying by using their knowledge of the language and by observing the situation in which the words were uttered.

In their earliest attempts at analysing two-word utterances, linguists followed the ‘Let’s pretend they’re talking Martian’ technique. Martin Braine (1963), of the University of California at Santa Barbara, listed all the twoword utterances produced by three 2-year-olds, Steven, Gregory and Andrew. The results were superficially puzzling. There were a number of inexplicable sequences such as MORE TAXI, ALLGONE SHOE, NO BED, BUNNY DO, IT DOGGIE. Such utterances could not be straight imitations, as it is unlikely that any adult ever said MORE TAXI, ALLGONE SHOE or BUNNY DO. Anyway, straight imitation would put too great a strain on the child’s memory. Braine counted over 2,500 different combinations uttered by one child. Are these then just accidental juxtapositions? Apparently not.

To his surprise, Braine noted that the combinations did
not
seem to be random. Certain words always occurred in a fixed place, and other words never occurred alone. Andrew, Steven and Gregory all seemed to have adopted a simple though genuine pattern when they put two words together. They had two distinct classes of word in their speech. One class contained a small number of words such as ALLGONE, MORE, THIS, NO. These words occurred frequently, never alone, and in a fixed position. They were labelled
pivots
, because the utterance appeared to pivot round them. The other class contained many more words which occurred less frequently, but in any position and sometimes alone. These words often coincided with adult nouns such as MILK, SHOE, BUNNY and so on. They are sometimes called
open
class words, since an ‘open’ class is a set of words which can be added to indefinitely.

For example, Steven always used WANT, GET, THERE, IT as pivots in first position, and DO as a pivot in second position. His open class words included a wide variety of names such as BABY, CAR, MAMA, DADA, BALL, DOLL, BUNNY, HORSIE. Steven seemed to have adopted a pattern which said, ‘A sentence consists of
either
a type 1 pivot followed by an open class word (P
¹
+ O),
or
an open class word followed by a type 2 pivot (O + P
²
)’:

Several other researchers who independently tried the exotic language technique confirmed this phenomenon by finding other children who formed word combinations in the same way as Andrew, Steven and Gregory (Brown and Fraser 1964; Miller and Ervin 1964). For a time, linguists were quite excited. They thought they might have discovered a universal first
grammar, a so-called
pivot grammar
. But, alas, disillusion gradually crept in. One by one, researchers noted that a number of children did not fit into this simple pattern. Although all children showed strong preferences for placing certain words in a particular position in an utterance, these preferences were not always strong enough to be regarded as genuine ‘rules’. In addition, some children used so-called pivots such as MORE, NO, by themselves, which disagreed with Braine’s finding that pivot words do not occur alone. And other children confused the picture by having pivot constructions as only a small portion of their total utterances.

Perhaps the biggest difficulty for pivot grammar was the appearance of utterances such as MUMMY SOCK, DADDY CAR, KITTY BALL, which occur in the speech of many children. Here two
open
class words seem to be juxtaposed, with no pivot in sight! Braine dismissed this problem by saying that O + O constructions were a second stage, which occurred only
after
the P + O and O + P phase. But this does not seem to be true of all children. Of course, there is nothing wrong with stating that some youngsters make sentences which can be P + O, O + P or O + O. It just does not tell us very much to say that ‘As well as pivot constructions, almost any other two words can occur together.’ But even if such empty statements were acceptable, it is not necessarily correct to assume that O + O utterances are random juxtapositions. There may be more reason behind them than appears at first sight, and the words may be related to one another in a highly structured way. It is quite inadequate to characterize a sentence such as DADDY CAR as O + O, since such a description cannot distinguish between several possible interpretations:

1 ‘Daddy is washing the car.’

2 ‘That’s Daddy’s car.’

3 ‘Daddy is under the car.’

Pivot grammars, therefore, were not as much use as was once hoped. They only described the rules used by a small number of children – or perhaps, more accurately, they characterized only a small portion of the output of most children. If one used pivot grammars in order to answer the question ‘Are two-word utterances structured?’, the answer would be: ‘Partially – children use pivot constructions but supplement them by apparently combining open class words at random.’

Disillusioned by the pivot grammars which resulted from the ‘Let’s pretend he’s talking Martian’ technique, later linguists tended to favour the second ‘Let’s guess what they’re trying to say’ approach. This is more time-consuming, since researchers must note not just the utterances themselves, but also the accompanying actions. Luckily, what young children say usually relates directly to what they do and see:

If an adult or an older child mounts a bicycle, there is no need for him to inform anyone who has seen him do it that he has done it. But a young child who mounts a tricycle will often ‘announce’ the fact:
I ride trike
!

(Bloom 1970: 9)

One of the first linguists to make a careful study of two-word utterances following this method was again Lois Bloom (1970, 1991; Bloom
et al.
1975). She kept a careful account of the actions accompanying the utterances of three children, Kathryn, Eric and Gia, and provided convincing interpretations of what they were trying to say. For example, it was quite clear what 21-month-old Kathryn meant on the two occasions when she uttered the words MOMMY SOCK. The first time, she said it as she picked up her mother’s sock, indicating that she meant ‘This is Mummy’s sock’. The second time was when her mother was putting Kathryn’s sock on Kathryn, so Kathryn was saying ‘Mummy is putting on my sock for me.’ Two-year-old Gia said LAMB EAR apparently meaning ‘That’s the lamb’s ear’ when her mother pointed to the ear on a toy lamb, and said, ‘What’s this?’ She said GIRL BALL when looking at a picture of a girl bouncing a ball, and presumably meant ‘The girl is bouncing a ball.’ She said FLY BLANKET when a fly settled on her blanket, probably meaning ‘There is a fly on my blanket.’

There is a possible objection to these interpretations. Is Bloom not reading too much into these utterances? Perhaps Gia was just saying ‘That is a lamb and an ear’, ‘That is a girl and a ball.’ ‘That is a fly and a blanket.’ Or perhaps she was just bringing up a ‘topic’ of conversation, and then making a ‘comment’ about it: ‘I’m talking about a
fly
, and it has involved itself with my
blanket
’, ‘I’m referring to a
girl
who is connected with a
ball
.’ This type of suggestion was first put forward in the mid-1960s to explain two-word utterances (Gruber 1967).