Do Classifiers Facilitate the Recognition of Relative Clause Structures in Chinese Elaine Andersen and Fuyun Wu Recent studies of relative clause (RC) processing in East Asian languages have provided conflicting evidence about the role played by noun classifiers. For example, research on the use of classifiers as a cue to construct a relative clause has yielded positive results for Japanese (Yoshida, Aoshima & Phillips, 2004), but not for Chinese (Hsu, Phillips & Yoshida, 2005). This difference may be purely a consequence of differences in Japanese and Chinese grammars, or they may be at least partially a result of methodological factors. One methodological consideration is that these studies presented their target sentences in isolation rather than embedded in discourse, which may have interfered with participants’ ability to utilize a classifier cue. The present study investigates whether classifiers can produce facilitation in Chinese when a more supportive discourse context is provided. In the first experiment, ten native Chinese speakers viewed pictures and listened to sentences while their eye movements were monitored. The last sentence on each trial was an instruction to point to a target picture The visual display was designed so that a restrictive relative clause was often necessary to indicate which of two pictures was being referred to. In the critical instructions, a mismatch between a classifier and the following noun provided an early cue to such an RC. Participants showed a small but reliable increase in early looks to the target picture following a mismatch (p<0.01), relative to a control condition in which the classifier and noun matched. In the second experiment, an additional ten Chinese speakers heard the same sentences, but with the passive marker ‘bei’ intervening between the classifier and the embedded noun. ‘Bei’ also serves as a cue to an upcoming relative clause, making the mismatch cue redundant. In this experiment, the advantage in the mismatch condition disappeared. A comparison of the current results with those of Hsu et al. (2005) suggests an interaction between lexical, syntactic, and discourse factors in on-line Chinese parsing. Relatively subtle cues to a RC (such as a mismatch between a classifier and a noun) have little observable effect when the context provides little reason to expect an RC to occur. The same cues can have a larger effect when discourse factors favor an RC. The difference between Japanese and Chinese may be in the relative weight of different parsing cues, due to the fact that Japanese is strictly head-final and Chinese is a mixed-order language. If comprehenders do not have a verb available until the end of the sentence, they may be more inclined to utilize other cues to upcoming structure, as the work on Japanese classifiers suggests.
|