| |
Abstract:
Reduced expressions such as the pronoun, he, play fundamental
roles in promoting discourse coherence by providing critical links
to integrate separate utterances into an existing model of
discourse (Gordon, Grosz & Gilliom, 1993; Grosz, Joshi &
Weinstein, 1995). This integration process has been found to be
modulated by the accessibility of the intended referent as
determined by its syntactic prominence and by the lexical features
of a reduced expression such as gender (Gordon, et al., 1993;
Gordon & Hendrick, 1998; Yang, Gordon, Hendrick, & Wu,
1999). We report five reading-time experiments in Chinese to
further systematically examine this integration process.
The use of Chinese provides a valuable opportunity in this
investigation because: First, Chinese has more flexible word
orderings than those in English. While English is predominated by
SVO word ordering, Chinese has OSV, SOV, and VOS word orderings.
This linguistic property of Chinese was employed to further inspect
how the syntactic and sequential structure of linguistic input
influences the mapping of discourse entities into the discourse
model. Second, Chinese is a pro-drop language with a "zero
pronoun', a phonologically null reduced expression that does not
exist in English. This linguistic property provided opportunity to
examine further the relative ability of two types of reduced
expressions to convey lexical information about the identity of
their referents in the processing of discourse integration.
In the current study, we systematically manipulated the
word-ordering structures and the forms of reduced expressions
(overt/zero pronoun) to examine how the comprehension of overt and
zero pronouns was modulated by the accessibility of the intended
referent as determined by the syntactic prominence of antecedent
referent, and the grammatically encoded features, particularly
gender, of reduced expressions. The result shows that the effect of
syntactic prominence determines the accessibility of discourse
referents from the model of discourse for later retrieval. They
have provided support for the characterization that the initial
steps in language comprehension are driven by syntactic and
sequential factors that influence how linguistic expressions are
mapped onto discourse representation (Gordon & Hendrick, 1998).
The results also indicates that a reduced expression is used as a
critical link to guide the integration of the semantics of the
predicate following the reduced expression with the associated
semantics of a preceding referent that matches its gender.
Different types of reduced expression induced different levels of
cognitive processes in such processing of semantics integration.
The implication of these results is discussed in the light of
centering theory (Grosz, et al., 1995) and discourse prominence
theory (Gordon & Hendrick, 1998). In sum, the current study
sheds light on our understanding of how sentences are integrated
dynamically into a model of discourse.
Gordon, P. C., & Hendrick, R. (1998). The representation and
processing of co-reference in discourse. Cognitive Science, 22,
389-424.
Gordon, P.C., Grosz, B. J., and Gilliom, L.A. (1993). Pronouns,
names, and the centering of attention in discourse. Cognitive
Science, 17, 311-347.
Grosz, B. J., Joshi, A. K., & Weinstein, S. (1995). Centering:
A framework for modelling the local coherence of discourse,
Computational Linguistics, 21, 203- 226.
Yang, C. L., Gordon, P. C., Hendrick, R., & Wu, J. T. (1999,
in press). Comprehension of referring expressions in Chinese.
Language and Cognitive Processes.
|
