ISSN : 1229-0718
In this study, we examined how the Contrast Principle was applied in a two-year olds longitudinal data which was tape recorded once a week for a year. The word substitution game, in which a word in a sentence was substituted for another word, the two or more words were contrasted keeping semantic relations with the predicate. The contrasted words kept the predictability with the predicate as subject, object or location. We noticed that the contrasted words which was in objective or locative relations were in a same semantic category. As to Keil's predictability tree, we analyzed the subject-contrasted words. Then we can suggest that in such early stage, children represent the similar ontological category as Keil's, and the M constraint are applied. We also could find the differentiation process as the first 4 months data were analyzed separately.