抄録
本稿では日本語の名詞複合語の知識の獲得を調査するために行った行動実験の調査報告をする。日本語を母語とする幼児24名(平均5歳8ヶ月)を対象に,意味判断課題を用いて3度に渡って実験を行った。結果,幼児の主要部理解はコントロールの大人の正答率よりも有意に低いものであることが明らかになった。これは,言語獲得の初期段階において幼児は非再帰的併合のみが利用可能であり,構造の内心性(あるいは投射と呼ばれる操作)はより後の段階において獲得されるという仮説と一致する。一方,幼児だけでなく大人も主要部理解の方が姉妹要素-主要部の意味関係理解の正答率よりも有意に高いものであったことから,幼児の意味関係に関する知識の獲得の遅れは,成熟ではなく統語論と談話のインターフェイスの複雑性に還元される可能性を論じた。
The present study reports on our longitudinal psycholinguistic experiment on the acquisition of noun compounds in Japanese. Twenty-four Japanese-speaking children (mean: 5 years and 8 months) were tested three times for comprehension of the Right-hand Head Rule and the Semantic Relation between the head and its sister element in noun compounds. The results show that children at this stage cannot fully project the head, which is consistent with the claim that the operation non-recursive Merge is only available to a child at the earliest stage of language development and the endocentricity of structures (or an operation called Projection) develops at later stages. In contrast, there is a significant difference in the percentage of correct answers between the Right-hand Head Rule and the Semantic Relation in the control adult group as well as in the children's group, which suggests that lateness in the children's acquisition of knowledge of the Semantic Relation cannot be reduced to maturation, and that the complexities of the (narrow) syntax-discourse interface also come into play.
