Phonological recoding in Chinese logograph recognition

Phonological recoding in a logographic script was examined in 4 experiments. Two factors were manipulated: orthographic and phonemic resemblance of the target to the correct word. The first 2 experiments, which used a verification task, showed that activation of the target's phonological code preceded the retrieval of its semantic code. When the target was masked (Experiment 2), the homophone effect was not dampened although accuracy was affected, suggesting that phonological activation is independent of lexical identification. The results of the idiom judgment task (Experiments 3 and 4) also supported the phonological recoding interpretation. It is argued that information about the logograph's overall configuration carried by the early-arriving low spatial frequency components allows the computation of a "prelexical" phonological code, although this information is not sufficient for lexical identification.