Chih-Hao Tsai's Research Page >

1995-06-01

Effects of Semantic Transparency and Morphological Structure On the Representation and Recognition of Chinese Disyllabic Words

Chih-Hao Tsai

University of Illinois at Urbana-Champaign

Tsai, C. H. (1995, June). Effects of semantic transparency and morphological structure on the representation and recognition of Chinese disyllabic words. Paper presented at the Joint Meeting of the Fourth International Conference on Chinese Linguistics and the Seventh North American Conference on Chinese Linguistics, Madison, Wisconsin.

Abstract

Two lexical decision experiments were carried out to investigate the representation and recognition of Chinese disyllabic words. Twenty-eight undergraduate students enrolled at National Chung- Cheng University at Taiwan participated in this study.

Experiment 1 found that the reaction times (RTs) for high frequency compounds were shorter than the RTs for low frequency ones. The RTs for compounds whose meanings are close to each constituent's dominant meaning (semantically transparent) were shorter than RTs for compounds whose meanings are not related to either constituent's dominant meaning (semantically opaque). This semantic transparency effect was larger in low word frequency condition than in high word frequency condition. The RTs for compounds composed of low frequency characters were shorter than the RTs for compounds composed of high frequency characters in the opaque, high word frequency and the transparent, low word frequency conditions (which resulted in a three-way interaction). Experiment 2 found that the RTs were much longer for the initial- opaque condition in which the dominant meaning of the initial character of a compound was unrelated to the meaning of the compound. The RTs for phonetic translations and loan words (PL words) were shorter than RTs for compounds.

These results have several implications on the representation and recognition of Chinese disyllabic words. First, high frequency words are stored in mental lexicon as independent lexical entries. Second, the lexical entries of low frequency compounds are not stored in mental lexicon. Third, the PL words are stored in mental lexicon as independent lexical entries, even for low frequency ones. Finally, these results provide evidence for a dual-process model that accounts for the recognition of Chinese disyllabic words.