語意關聯與類別語意關係對兒童中文語意發展影響的縱貫式研究

A Longitudinal Study of Semantic Association and Categorical Relatedness on Children's Semantic Processing of Chinese Characters

過去中文語意發展的研究，皆採用橫斷式取向觀察年齡變化的差異，而且沒有將語意關聯與語意關係對於語意處理的影響清楚區分。語意關聯係指字對是以自由聯想建立的關聯性字對；語意關係則是指字對為同類別語意概念。本研究採縱貫式取向，正交地操弄語意關聯（高、低）與類別語意關係（高、低）兩個變項，進一步檢驗11歲兒童語意知識的發展，探討這兩個變項對語意處理的貢獻。實驗以語意判斷作業進行，24位兒童判斷依序出現的兩個中文字是否有意義關係，間隔兩年再次進行相同作業。本研究有兩項主要發現：（一）時間變項主要效果顯著，兒童第二次（兩年後）的判斷正確率優於第一次參與；（二）語意關聯與類別語意關係的交互作用顯著，高低類別語意關係的差異，在高語意關聯時達顯著，低語意關聯時則否；高低語意關聯的差異，在高類別語意關係達顯著，低類別語意關係時則否。推論：（一）隨著年齡成長與兩年的教育學習經驗，兒童對於語意表徵的掌握能力更加精緻化，能以抽象思考能力整合不同層級的類別語意關係；（二）語意關聯與類別語意關係的交互促進效果，在語意關聯與類別語意關係都有增強，穩定的語意促進效果是建立在字彙同時具有高語意關聯及高類別語意關係的組織方式上。

Previous developmental studies with regard to semantic processing used cross-sectional approaches to examine developmental changes and did not make distinctions between semantic association and categorical relatedness. Semantic association was defined as character pairs produced by free association tasks; categorical relatedness was referred to character pairs within the same semantic category. The present study used a longitudinal approach to further investigate the roles of semantic association and categorical relatedness in the developmental changes of semantic processing. We orthogonally manipulated semantic association (strong, weak) and categorical relatedness (high, low) to examine developmental changes of 11-year-old children in a two-year interval. Children were asked to decide whether two visually-presented Chinese characters were semantically related and tested again two years later. There were two main results. First, the main effect of time was significant. Children performed better on accuracy of the second measure as compared with the first measure. Second, we found an interaction between semantic association and categorical relatedness on accuracy. The difference between the high and low categorical relatedness was significant for the strong semantic association, not for the weak semantic association. The difference between the strong and weak semantic association was significant for the high categorical relatedness, not for the low categorical relatedness. There are two implications in the present study. First, as age increased, accumulated learning experiences may enhance children to build more elaborate semantic representations. They may use more abstract manners to integrate different categorical relationships. Second, the interaction of semantic association and categorical relatedness suggests that semantic knowledge may be organized by both semantic association and categorical relatedness. The stable semantic enhancement is based on characters with strong semantic association and high categorical relatedness.