Content area
The aim of this thesis is to investigate whether knowledge of semantic radicals influences Chinese character identification. A series of three investigations addressed the effects of semantic radical properties on the time course of character processing during: (1) orthographic processing by using a lexical decision task manipulating semantic radical combinality (neighbourhood density); (2) semantic processing by manipulating semantic radical consistency and character transparency using a semantic categorisation task; and (3) the effect of semantic radical characteristics on name production by examining the extent of semantic processing when character orthography is controlled in a picture-word interference task.
These investigations were conducted with native Mandarin speakers residing in the United Kingdom using behavioural and event-related potential methods. Results suggest that: (1) semantic radical combinability has an effect on post lexical processing; (2) effects of semantic radical consistency are limited to opaque characters at an early stage of lexical access specifically at the recognition potential and during post lexical processing at the N400; and (3) semantic radical effects are task dependent and rely on depth of semantic processing.
Findings are discussed with respect to existing models of Chinese character recognition. The data supports multi-level models of orthographic-semantic processing in Chinese that assume two kinds of representation: a radical level and a character level. It is argued, however, that the present findings challenge these models, as they cannot account for observed interactions between the effects of semantic radical consistency and character transparency. To explain the present results, cognitive models of Chinese character identification need to assume direct mappings between semantic radicals and the semantic system.