Statistical Learning as a Key to Cracking Chinese Orthographic Codes

This study examines statistical learning as a mechanism for Chinese orthographic learning among children in Grades 3–5. Using an artificial orthography, children were repeatedly exposed to positional, phonetic, and semantic regularities of radicals. Children showed statistical learning of all three regularities. Regularities’ levels of consistency impacted character recognition (better performance with high positional consistency; Experiment 1) and generalization (with graded effects of semantic consistency; Experiment 3), though level of phonetic consistency did not affect performance (Experiment 2). Results support a connectionist model of Chinese orthographic learning that emphasizes statistical learning as a powerful tool to crack Chinese orthographic codes.