How the Chinese writing system can reveal the fundamentals of hierarchical lexical structure

Abstract

Compounding is the most common way in which languages can create new words from existing ones. The investigation of how compound words are represented and processed in the mind can therefore reveal fundamental properties of lexical representation and processing. Because of the properties of its morphology and its writing system, Chinese offers a special opportunity for such investigation. In Chinese, it is common for compound words to contain three or more constituents, which are organized into hierarchical structures. Using a lexical decision task, we examined whether the arrangement of such sublexical hierarchies within triconstituent compound words would affect their ease of visual processing. We found that left-branching structures were processed more accurately and more quickly than right-branching ones (left-branching triconstituent compounds are those in which an embedded two-character compound is the modifier and a single character is as the head, as in the English left-branching structure hot dog bun). This processing advantage of left-branching structures was supported by the results of a metalinguistic segmentation task in which participants showed left-branching preferences when presented with special triconstituent compounds which have no feasible branching structure (loanwords and flat structure compounds). In the Chinese character-based writing system, left-branching and right-branching compounds are written exactly the same way. This is in contrast to English in which it is common for the branching direction of a triconstituent compound to be indicated by the presence or absence of a space within the compound structure (e.g., watchdog group, consumer watchdog). Thus, the results from Chinese support the view that hierarchical constituent structure plays a role in processing in the absence of orthographic support and that the observed left-branching preference reflects central aspects of lexical representation and processing.