A behavioral and ERP study of the role of radical position distribution in character recognition : evidence for position-specific radical representation

Abstract

The present study investigated the role of radical position distribution in character recognition using a masked primed lexical decision task. Thirty-four normal adults (mean age: 22 years old) who were skilled readers of Chinese and native Cantonese speakers participated in the study. Radicals were used as the primes and complex characters were used as the targets. Participants were asked to judge whether the complex characters presented is a real character or not as quickly and accurately as possible. Response latencies, accuracy rates and ERP data were recorded. Behavioral data indicated that characters with position-biased related primes were recognized significantly faster than those with position-biased unrelated ones and the reverse was observed in position-neutral radical priming. ERP data only observed lexicality effect in N400 component. The findings have demonstrated the complex role of radical position distribution in character recognition and substantiated the argument for position-specific radical representation.