Word order, animacy, and agreement cues in sentence processing by Li Mandarin EFL learners

• Main Author: Lin, Yowyu
• Other Authors: Doughty, Catherine
• Published: University of Hawaii at Manoa 2009
• Online Access: http://hdl.handle.net/10125/7109

Research within the Competition Model has shown that different language speakers use different strategies to interpret sentences. Despite the fact that crosslinguistic studies have incorporated a variety of cues in the experiments, studies investigating Mandarin used mostly word order and animacy cues only. Modifying experiment designs of the previous studies and adding agreement cues in the research, the present study presented subjects a series of both grammatical and ungrammatical English sentences and asked them to identify the "actor" of these sentences. One group of English native speakers and two groups of Taiwanese English learners with different English proficiency were recruited. Three research questions were investigated in the present study: (1) Which cues, word order, agreement and animacy, would be used most for the three different groups of subjects? (2) What kind of language transfer would be found in nonnative learners of English, especially Taiwanese high school students? (3) Would learners of English with different levels of proficiency differ in their use of cues in sentence processing? Findings of the present research indicated that English native speakers used word order as the primary cues, while intermediate and advanced Taiwanese English learners used mainly animacy and word order cues respectively, which demonstrated that a different length of exposure to a foreign language did influence learners' strategies in sentence processing. In addition, results of the present study also revealed a pattern of forward transfer in sentence processing by intermediate nonnative subjects, while the advanced group of subjects did not show this transfer. Interestingly, after modifying previous research designs, the effect of animacy cues in the group of English native speakers did not reach significance in the present study. This may shed light on the influence of real world bias in the previous studies. === xv, 133 leaves