Article details

Research area
Natural language & AI

Proceedings of the 1986 National Conference on Artificial Intelligence (AAAI 86)


J. McDowell

Using commonsense knowledge to disambiguate prepositional phrase modifiers


This paper describes a method of using commonsense knowledge for discarding spurious syntactic ambiguities introduced by post-verbal prepositional phrase attachment during parsing. A completely naïve parser will generate three parses for sentences of the form NP-V-Det-N-PP. The prepositions alone are insufficiently precise in meaning to guide selection among competing parses. The method is imbedded in the Kind Types System (KT) which employs commonsense knowledge concepts, including prototype and inherent features (generic information) and ontological classifications. The generic information is drawn from published psycholinguistic studies on how average people typically view the world. This method is employed in preference strategies which appeal to the meaning of the preposition combined with information about the verbs and nouns associated with it drawn from the text and from the generic and ontological databases. These determine which syntactic structures generated by a semantically naïve parser are commonsensically plausible. The method was successful in 93% of cases tested.

Read/download now