Many have claimed that children treat relative adjectives "nominally." Studies have shown that children refuse to relabel objects, use relative terms only when there are extreme differences between objects, and ignore the relative differences of objects within categories. Although none of these reported uses are truly nominal since they imply relativistic judgments, they do suggest limits on children's understanding of relative terms. However, there is little evidence on which, if any, of these uses predominates. In 4 experiments, we examined 2-, 3-, and 4-year-olds' abilities to use big and little. We found that interpretations in accord with the proximate object of comparison dominated the children's uses overall. The "errors" that the youngest children did make primarily reflected an object's position relative to the stimulus array as a whole. The shifting from one standard to another and the magnitude of difference between objects had minor effects on performance. We also found that the positive term, big, was used more flexibly than little. Our overall pattern of results suggests that children's so-called nominal errors reflect the acquisition of the relativistic ways in which the terms are used by adults.
Bibliographical noteFunding Information:
Children's performances across a variety of tasks suggest that they do not fully understand relative adjectives such as big, long, little, tall, short, high, and low until they are 4 or 5 years of age (Donaldson & Wales, 1970; Ehri, 1976; Klahr & Siegler, 1978; Maratsos, 1973; Sinclair-de-Zwart, 1969; Trabasso, 1977). Children's difficulty with relative terms is not surprising; relative terms unlike the more rapidly acquired nominal terms (MacNamara, 1982; Nelson, 1973), do not refer to stable attributes of objects but instead refer to a relationship between an object and some frequently updated standard of comparison. A single object could be labeled, for example, as big in one context and little in another if the Portions of this research were presented at the meetings of the Society for Research in Child Development, Toronto, 1985. This research was supported by National Science Foundation Grant No. BNS 81 09888, National Institute of Health Grant No. PHS ROi HDI9499, and Public Health Service Grant No. KO4 HD589 01. We gratefully acknowledge the assistance of Carol McCord, Mary Jo Ratterman, Andrew Brown, and Beverly Spangler for their help in data collection;c hildren, parents, and staff of Cherry Hill Day Care for their participation; and Judith Johnston, Susan S. Jones, and Bea Gattuso for their helpful comments on the manuscript.