Semantic Interpretation as Higher-Order Deduction

Traditional accounts of the semantic interpretation of quantified phrases and its interaction with reference and ellipsis have relied on formal manipulations of logical forms (quantifier raising) or complicated denotations for phrases (Cooper storage). These complications are one of the motivations for the development of categorial semantic theories. However, these theories appear to face their own difficulties in accounting for the full range of scoping possibilities in natural language, and have been losing some of their original elegance. Experiments suggest that the insights of the traditional theories may be revived in a semantically more respectable form by taking advantage of the abstraction mechanisms of higher-order hereditary Harrop formulae, as implemented in Prolog, to represent scoping and extraction dependencies without the need for formal conditions on logical forms or elaborate denotations for phrases. The technique used is an adaptation of a method of Felty and Miller...

• Type: Conference
• Year: 1990
• Where: JELIA
• Authors: Fernando C. N. Pereira
• Keywords: Artificial Intelligence, Categorial Semantics, JELIA 1990, Logical Forms, Semantic Interpretation