A Computational Model of Referring

We present a model of referring within the framework of a general theory of speech acts and rationality. Following an act of referring, the hearer is expected to accomplish two tasks: first, recognize the speaker's literal goal, and second, recognize and satisfy the identification constraints. We show how the literal goal of referring is achieved, and how referent identification is possible without the formal device of standard names. Furthermore, we show how referring expressions can be used to inform and make requests. This theory has application to the design of a natural-language utterance planning system that formulates utterances to achieve its goals.