COMPUTATIONAL INTERPRETATIONS OF THE GRICEAN MAXIMS IN THE GENERATIONOF REFERRING EXPRESSIONS

We examine the problem of generating definite noun phrases that are ap propriate referring expressions; that is, noun phrases that (a) succes sfully identify the intended referent to the hearer whilst (b) not con veying to him or her any false conversational implicatures (Grice, 197 5). We review several possible computational interpretations of the co nversational implicature maxims, with different computational costs, a nd argue that the simplest may be the best, because it seems to be clo sest to what human speakers do. We describe our recommended algorithm in detail, along with a specification of the resources a host system m ust provide in order to make use of the algorithm, and an implementati on used in the natural language generation component of the IDAS syste m.