The anecdotes that parents retell about their child’s early speech (to the intense embarrassment of the grown-up child) usually involve examples of the strange use First language acquisition 179 of words. Having been warned that flies bring germs into the house, one child was asked what “germs” were and the answer was “something the flies play with.” It is not always possible to determine so precisely the meanings that children attach to the words they use.

It seems that during the holophrastic stage many children use their limited vocabulary to refer to a large number of unrelated objects. One child first used bow-wow to refer to a dog and then to a fur piece with glass eyes, a set of cufflinks and even a bath thermometer. The word bow-wow seemed to have a meaning like “object with shiny bits.” Other children often extend bow-wow to refer to cats, cows and horses.

This process is called overextension and the most common pattern is for the child to overextend the meaning of a word on the basis of similarities of shape, sound and size, and, to a lesser extent, movement and texture. Thus the word ball is extended to all kinds of round objects, including a lampshade, a doorknob and the moon. Or, a ticktock is initially used for a watch, but can also be used for a bathroom scale with a round dial. On the basis of size, presumably, the word fly was first used for the insect and then came to be used for specks of dirt and even crumbs of bread. Apparently due to similarities of texture, the expression sizo was first used by one child for scissors, and then extended to all metal objects. The semantic development in a child’s use of words is usually a process of overextension initially, followed by a gradual process of narrowing down the application of each term as more words are learned.

Although overextension has been well-documented in children’s speech production, it isn’t necessarily used in speech comprehension. One two-year-old used apple, in speaking, to refer to a number of other round objects like a tomato and a ball, but had no difficulty picking out the apple, when asked, from a set of round objects including a ball and a tomato.

One interesting feature of the young child’s semantics is the way certain lexical relations are treated. In terms of hyponymy, the child will almost always use the “middle”-level term in a hyponymous set such as animal – dog – poodle. It would seem more logical to learn the most general term (animal), but all evidence indicates that children first use dog with an overextended meaning close to the meaning of “animal.” This may be connected to a similar tendency in adults, when talking to young children, to refer to flowers (not the more general plants, or the more specific tulips).

It also seems that antonymous relations are acquired fairly late (after the age of five). In one study, a large number of kindergarten children pointed to the same heavily laden apple tree when asked Which tree has more apples? and also when asked Which tree has less apples?. They just seem to think the correct response will be the larger one, disregarding the difference between more and less. The distinctions between a number of other pairs such as before/after and buy/sell also seem to be later acquisitions.

Despite the fact that the child is still to acquire a large number of other aspects of his or her first language through the later years of childhood, it is normally assumed that, by the age of five, the child has completed the greater part of the basic language acquisition process. According to some, the child is then in a good position to start learning a second (or foreign) language. However, most people don’t start trying to learn another language until much later. The question that always arises is: if first language acquisition was so straightforward and largely automatic, why is learning a second language so difficult?