In this paper I will first show that a purely syntactic condition, closely related to the Binding Theory, imposes a restriction on the interpretation of definite descriptions. This restriction has, to my knowledge, not been observed before. Next, I will make a proposal as to how this condition can best be incorporated in a grammar model. I suggest that the condition can be naturally incorporated in the rules for interpreting syntactic discourse models. The proposal has additional advantages for the analysis of natural language, making it possible to avoid unnecessary and defer real ambiguities.
|Number of pages||10|
|Journal||IPO Annual Progress Report|
|Publication status||Published - 1995|