Structuring and retrieval of the complex predicate arguments proper to the NKRL conceptual language

Abstract

Extracting a “reasonable” amount of the information contained in “narrative” documents like, e.g., news agency stories, requires the use of a reasonably rich representation language, which in turn requires a query languages powerful enough to exploit this richness. This paper deals with the conceptual structures used in NKRL (NKRL = Narrative Knowledge Representation Language) to describe the “structured arguments” of the NKRL semantic predicates. We describe here the query language, and the corresponding matching algorithm, which allows us to retrieve correctly the information stored in these arguments.