Introduction

Children have learned over 14,000 words by age 6. It is estimated that children acquire about nine new words a day from the age of 18 months to 6 years (Carey, 1978). How do they acquire so many words at such a rapid rate? If children's acquisition of words involves an inductive process, i.e. unbiased hypothesis testing, then, their fast learning at an earlier stage is difficult to explain. Consider an example of a child learning a new word. A mother points to a rabbit and says ‘rabbit’ to her child. There seem unlimited possibilities that the child might guess the word is intended to describe. It could refer to the whole object (the rabbit), a part of the object (long ears, a tail, red eyes, a leg, etc.), the substance the object is made of (fur, bones, flesh, etc.), some property of the object (furry, white, dirty, etc.), and so on. In spite of an infinite set of possibilities about the novel word's meaning, the child rapidly and successfully connects the novel word to its meaning. This is the well-known inductive problem that Quine (1960) has posed. As a solution for this problem, many proposals have assumed that children might be equipped with certain implicit biases or constraints that narrow down the possibilities when ascribing meaning to a novel word (E. Clark, 1987; Landau, Smith & Jones, 1988; Markman, 1994; Soja, Carey & Spelke, 1991).