Experimental investigations on children's early structural representation: a view from classifier phrases in Mandarin

Linguistic theories divide on whether children's early representation of language is based on hierarchical structural relations or on linear relations. To shed further light on this debate, the present study investigated Mandarin-speaking children's choices of classifiers for noun-noun compounds, where the first noun is a modifier noun, and the second noun is a head noun. In these noun-noun compounds, the classifier agrees with the head noun, rather than the linearly closer modifier noun. Two experiments were conducted. Experiment 1 explored whether 4- to 6-year-olds would choose the classifier for the compounds based on linear distance or structural relations. Experiment 2 examined whether children's choices were influenced by their lexical knowledge of specific classifiers. The results of Experiment 1 show that 4-, 5- and 6-year-olds all chose classifiers that were compatible with head nouns, favoring the structural relation, but the accuracy rate of the 4-year-olds was significantly lower than that of the 5- and 6-year-olds. The findings of Experiment 2 indicate that once the 4-year-olds' lexical knowledge of specific classifiers was improved, their accuracy rates also significantly improved. The findings provide new evidence for the proposal that children's early representation of language is based on hierarchical structural relations rather than linear relations.