Which words do children learn earliest, and why? These questions bear on how humans organize the world into semantic concepts, and how children acquire this parsing. A useful perspective is to think of how bits of experience are conflated into the same concept. One possibility is that children are born with the set of conceptual conflations that figures in human language. But assuming (as we will) that most semantic concepts are learned, not innate, there remain two possibilities. First, aspects of perceptual experience could form inevitable conflations that are conceptualized and lexicalized as unified concepts.

In this case, we would have cognitive dominance: concepts arise from the cognitive-perceptual sphere and are simply named by language. A second possibility is guistic dominancee world presents perceptual bits whose clumping is not pre-ordained, and language has a say in how the bits get conflated into concepts. We propose that both cognitive and linguistic dominance apply, but to different degrees for different kinds of words (Gentner 1981, 1982). Some bits of experience naturally form themselves into inevitable (preindividuated) concepts, while other bits are able to enter into several different possible combinations.

Relational relativity and the division of dominance

Embracing both cognitive and linguistic dominance may seem to be a vague middle-of-the-road position. But we can make the distinction sharper by asking which applies when. We suggest a larger pattern, a division of dominance (Gentner 1988). The Division of Dominance continuum is shown in figure 8.1. At one extreme, concrete nouns - terms for objects and animate beings - follow cognitive- perceptual dominance.