An Integrated Approach to Anaphora
【摘要】：In recent years, the field of linguistic study has been witnessing strong interests in the anaphoric study. With a substantial amount of empirical and theoretical research in this field, the anaphoric study has yielded many convincing research findings. However, the role that knowledge-based inference plays in antecedent determination remains an unexplored area. Based on the previous study on anaphora and especially the recent cognitive research in this aspect, the present paper, combining Reference Point Theory, Discourse Representation Structure and borrowing Ontology-based inference from Artificial Intelligence, establishes an integrated solution to antecedent determination. The model enjoys robust and convincing explanation for many kinds of anaphora, even including associative anaphora. The implications for Natural Language Processing are also suggested.