Abstract
Japanese two-kanji compound words (Jukugo) have two sets of neighbors: Front-neighbors that share the same front-kanji and Rear-neighbors that share the same rear-kanji of the Jukugo. The purpose of the present study is to investigate the factors affecting semantic similarity among Jukugo neighbors. In a neighbor recall-task using Kyoiku kanji, participants were presented with ten front-kanji and ten rear-kanji on a questionnaire. The participants were asked to write down a rear-kanji and front-kanji to make a neighbor up to eight neighbors within a minute for each cue kanji. The results were analyzed in terms of how factors such as number of neighbors and recall frequency of neighbors affect the semantic similarity rating values among neighbors.
