THE ON-LINE PROCESSING AND ANTICIPATION BUILDING IN NATIVE MANDARIN SPEAKERS AND LATE DUTCH-MANDARIN LEARNERS

Abstract

In order to explore late L2ers’ on-line processing of mass/count syntactic cues which are unique-to-L2 constructions and can only be acquired through implicit learning, a Visual World Paradigm experiment and a Reading for Comprehension experiment were conducted on high proficiency late Dutch-Mandarin learners and native Mandarin speakers. The mass/count syntactic cues in Mandarin (the Adj-CL word order, and the insertion of the modification maker de after a classifier) are unique-to-L2 constructions for Dutch-Mandarin learners, since there exists no classifier in Dutch. Li, Barner, & Huang (2008) is the first experimental study explored native Mandarin speakers’ off-line using of the mass/count syntactic cues. They found that nominal phrases with different mass/count syntactic cues led native Mandarin speakers to have different interpretations. It remains unclear how native Mandarin speakers on-line process the mass/count syntactic cues, and whether late L2-Mandarin learners can acquire these mass/count syntactic cues and exhibit native-like behaviours in real time processing. To tackle these questions, the current research tested native Mandarin speakers’ and high proficiency late Dutch-Mandarin learners’ predictive processing of nominal phrases with different mass/count syntactic cues. Both of the two syntactic cues (the Adj-CL word order, and the insertion of the modification maker de after a classifier) were used in the current research, as well as typical count and mass nouns. The results showed that native Mandarin speakers can take advantage of the mass/count syntactic cues in real time predictive processing. Late Dutch-Mandarin learners exhibited native-like behaviours in the Visual World Paradigm experiment, but non-native-like behaviours in the Reading for Comprehension experiment. The findings indicated that late L2ers can acquire unique-to-L2 constructions through implicit learning, and their processing difficulties are caused by limited cognitive resources but not deficient representations.