摘要： 基于文[3]所述的理解系统为一个中文段落生成的理解图(由事件及其意义连贯关系组成),本文提出了一个汉语对话模型。我们首先研究了疑问句的含义及其事件表示法(Q_s=(q,E_q),然后研究了问句和理解图中可能描述的答案之间的形式关系(φ:Q→2~(SR+)+U2~(CR)U{是/否},从而形成了在[3]提出的理解模型基础上的有关根据问句寻求答案的对话模型。

Abstract: We, first, research the meaning of an interrogation and its event expression (Qn = (q, Eq)) and then, get the formal relationships ( : Q→2SR+ ∪ 2CR∪ {Yes/No}) between the interrogation and the possible answer described in the understanding graph so as to form the dialogue model that looks for the answer according to the interrogation on the understanding model proposed in [3].