Parsing for Role and Reference Grammar

Parsing of human languages is a well studied subject, but most algorithms will handle either fixed word order languages or free word order languages. In this paper we describe extensions to a chart parser in order to enable varying degrees of word order flexibility and parsing via templates instead of rules. We have chosen to use templates instead of rules because this enables more information to be captured, which in turn will make it easier to extract the meaning from the sentence. The grammatical paradigm used in this paper is Role and Reference Grammar, but we expect the methods to be useful for other grammatical paradigms.

Published online: 21 November 2008
https://doi.org/10.1075/slcs.105.30gue