Perception, Knowledge and Young Korean Children's Word-Learning

Abstract

A novel word said to a child could in principle have countless number of meanings. Thus, it would be logically impossible for a child to learn a single word if she/he considers all the possibilities to figure out the correct meaning. This logic of word-learning problem led workers to propose special constraints or principles that could help young childrens word learning (see Bloom, 1995, 2000; Landau, 2000). Two of such constraints considered in this study are shape bias (Landau, 2000; Landau, Smith, & Jones, 1988, 1992, 1998) and ontological constraint (Soja, Carey, and Spelke, 1991, 1992). Shape bias states that young childrens initial interpretation of novel names respect the shape of objects, and thus, at the initial stage of word learning, the shape image (among other perceptual images) of an object is most likely mapped onto the sound form of the word. Ontological constraint states that childrens pre-linguistic knowledge about ontological distinctions between solid objects and non-solid substances, among others, constrains the initial interpretation of word meanings. Evidence supporting these claims is substantial (see Bloom, 2000; Landau, 2000), suggesting that both constraints are used by young word-learners. What is controversial, however, is the true nature of the shape bias; is shape bias found because shape is actually respected in word-learning or only because shape is a source of information about object kind. Soja, et al. (1992) prefer the latter reason over the former reason which is opted for by Landau et al. (1988, 1992, 1998).