2026/07/24 by Mengxia Chen, Katsuo Tamaoka
Neuroscience · Psychology · #Comprehension #Confirmatory factor analysis #Construct (python library) #Kanji #Latent semantic analysis #Neurobiology of Language and Bilingualism #Reading and Literacy Development #Reading comprehension #Second Language Acquisition and Learning #Similarity (geometry) #Structural equation modeling
paper · pdf · doi:10.1016/j.actpsy.2026.107460
published in Acta Psychologica 269, 107460 (Elsevier BV)
openalex publication_date 2026/07/24 · openalex created_date 2026/07/25 · openalex updated_date 2026/07/26
Chinese learners of Japanese as a foreign language (JFL) are often assumed to benefit from first-language kanji knowledge when reading Japanese, but this contribution has seldom been tested within an explicit causal framework. Here, we examine how kanji-based lexical knowledge supports Japanese text comprehension in Chinese JFL learners. Two-kanji words were categorized into three types based on orthographic and semantic similarity between Chinese and Japanese (i.e., Identical, Partially Overlapping, and Unrelated) and confirmatory factor analysis showed that all three types significantly loaded onto a single latent construct of kanji word knowledge. Structural equation modeling (SEM) further revealed that kanji word knowledge strongly predicts text comprehension, both indirectly through its facilitative effect on general lexical knowledge and directly as an independent contributor. These findings extend previous research by demonstrating, within a unified causal model, the central and robust role of kanji-based lexical knowledge in Japanese text comprehension among Chinese JFL learners.