Logographeme type and token frequency effects during Chinese character recognition : an event-related potential study

Abstract

This study aims to provide evidence supporting the existence of logographeme as a unit of representation during Chinese character recognition. Radicals have long been considered a basic unit in Chinese character recognition. However, in studies of dysgraphic patients, writing errors in characters could not be regarded as radical level errors but were considered logographeme level substitution errors (Law & Leung, 2000; Han, Zhang, Shu & Bi, 2007). To assess whether logographeme also affect the recognition process as found in writing studies, logographeme type frequency (high versus low) and logographeme token frequency (high versus low) were investigated using an event-related potential (ERP) lexical decision task. An interaction between logographeme type and token frequency was found at the occipital N170 component with greater negativity in high type-low token frequency and low type-high token frequency logographemes. Low logographeme type frequency also showed greater negativity at the N400 component. No logographeme effects or interactions were found in the P100 and P200 components. This study provides evidence supporting the existence of logographeme as a unit in Chinese character recognition and that logographemes are activated during orthographic processing stage and perhaps during lexical-semantic retrieval.