How Children Learn What Words Mean

Children learn language effortlessly and completely voluntarily. They learn new words miraculously fast. A teenager masters about 60,000 words of their mother tongue by the time they finish high school. In speech and writing, they may not actively use more than a limited subset of the words, but they understand them. A simple calculation reveals that they have learned an average of 9-10 words per day during childhood. A single example of how a word is used is often sufficient for learning its meaning. Despite the incredible efficiency of language learning, the underlying mechanism is still largely unknown.

How do children know what a word refers to? When a new word is introduced, there is often a multitude of objects, features of objects, and actions in the situation that could be the meaning of the word. There are many cues that can help to select the right meaning. A first cue is that words do not mean just anything. Our minds perceive different kinds of structures in the world and the meanings, to a large extent, follow these structures. Above all, there are things in the world that we need to talk about. The things come in categories and we have words for them that have meanings that correspond to these categories.

In my book, The Geometry of Meaning, I have put forward the thesis that children do not learn words in isolation, but they sort them into meaning domains. Examples of such domains are animals, fruits, colors, and vehicles. Once they start learning words in certain domains, they can easily pick up new ones. Thus, if a three-year-old knows "red," "yellow," "blue," and "green" in the color domain, they can easily learn "purple," "turquoise," and "beige" after just a couple of examples.

In contrast, if a child's parents are in financial trouble, the child may often hear words like "bills," "bank loans," and "mortgage." The child cannot learn what these words mean until they have acquired the corresponding concepts and this in turn requires the child to understand the financial domain. Children may be able to understand "buying" and "selling" as concrete actions, but this does not mean that they understand the concept of "money" (one child described their mother's actions at the ATM as "mum buying money"). This is even more difficult now that coins and notes are disappearing from everyday life. While language can help children learn concepts, it is not always enough. Try explaining the meaning of the word "mortgage" to a three-year-old!

I, therefore, formulated a general establishment thesis concerning semantic domains: If one word from a domain is learned during a certain period, then other (common) words from the same domain should be learned during roughly the same period.

In order to test the thesis, I have analyzed data from the CHILDES corpus of children’s spoken language. The publicly available web-based ChildFreq application is a highly efficient tool for such investigations. In the analysis of child language data, the establishment of a word in the vocabulary of children was analyzed in terms of the average frequency of the word’s usage at a certain age. Words often start with zero usage, increase rapidly, and then level off once they are established in a child's vocabulary. I call the period of most rapid increase the establishment period for a word. Between 12 and 24 months of age, children experience a spurt in which they learn many words for ordinary objects, while color words are acquired a little later.

However, some domains take longer to establish. One such example is the domain relating to life and death. The establishment of the words live, die, alive, and dead takes place between 30 and 42 months of age.

Many words, in particular adjectives, can be described in terms of dimensions. For example, the vertical dimension can be described by the opposites "high" and "low" for objects, such as a chair or a building, or "tall" and "short" for people. This dimension is also central to the meaning of the verbs "lift" and "lower." Or take the dimension of temperature, described by the word pair "hot" and "cold." Another example is "good" and "bad" which stands for a dimension of evaluation.

Another test of the establishment hypothesis studied how children learn adjectives that refer to the same domain. Together with speech therapist colleagues from Lund University, I conducted an experiment in which we checked how well children knew the meaning of opposite words. Three- and five-year-olds were asked a series of questions such as "If a thing is not cold, then it is...?" where the child had to fill in the missing adjective.

For the most part, the method worked well, with children either giving a correct answer or not being able to answer. Experiments always involve surprises: A child who was asked "If a person is not old, then he is ...?" answered " ... he is not so wrinkled in the face." There was a clear difference between the pairs of words mastered by the three-year-olds compared to the five-year-olds. The three-year-olds mastered basic domains like size and age, but not more abstract ones like "cheap"/"expensive" and "fast"/"slow," which the five-year-olds did. The most important result of the experiment was that if a child masters one word in a pair of opposites, there is a high probability that they will also master the other word. This supports my thesis that children learn the common words in a semantic domain at approximately the same period of language development.

Although there are still many enigmas surrounding how children can learn the meanings of new words so quickly, the insight that they do not learn words one by one, but rather domain by domain, partly explains the language learning miracle.

Facebook/LinkedIn image: Monkey Business Images/Shutterstock