幼児が拗音表記(e.g.,きゃ)の読みを習得するのに,/Ci/+/jV/→/CjV/という混成規則を利用しているのか,機械的対連合の方略を用いているのかを3つの研究により検討した。研究1では,5-6歳の保育園児を対象に,21個の清拗音表記の習得状況の調査を行った。その結果,習得数を横軸,習得人数を縦軸にした分布図は習得数0個と14個以上に分かれるU字型の分布となり,習得が開始から短時間で上限近くまで進行することが示唆された。研究2では,習得途上の子どもに創作拗音表記(Xゅ/dju/)を提示し,対象児の8割がこれを正しく読めることが示され,習得途上の子どもが混成規則を利用していることが示唆された。研究3では,混成規則を明示的に教授し,6割の子どもがそれを新規の事例に適用できることが示された。本研究の結果は,幼児の拗音表記の読み習得において混成規則の利用が中心的役割を果たしていることを示唆している。

The Japanese language uses a syllabary, rather than an alphabet. However, 21 of the /CjV/s are not represented by a single syllable, but rather by combinations called yo-on combinations, in which /Ci/ is subscripted by /ja/, /ju/, or /jo/ (e.g.,きゃ=/kja/). The present study aimed to examine whether young Japanese children utilize the blending rule (i.e., /Ci/+/jV/→/CjV/) when learn to read the combinations or learn them by rote. In Study 1, 5- to 6-year-old children took a reading test. The results indicated that most of the children were divided between those who could not read the combinations at all and those who could read more than 14 combinations. This suggests that combinations are mastered in a small amount of time. In Study 2, 5- to 6-year-old children who could read between 11 and 20 combinations were asked to read a pseudo yo-on combination X+ゅ(/dju/). The results indicated that 80% of them could read it. In Study 3, 5- to 6-year-old children who could read no yo-on combinations were instructed in the blending rule by showing them some examples. It was found that 60% of them could apply to rule to other items. The present studies suggest that the use of the blending rule plays a central role in learning to read yo-on combinations.