A Linguistically Grounded Annotation Language for Spatial Information

James Pustejovsky, Jessica Moszkowicz, Marc Verhagen
Department of Computer Science
Brandeis University
415 South Street
Waltham, MA


An understanding of spatial information in natural language is necessary for many computational linguistics and artificial intelligence applications. In this paper, we discuss what problems researchers face with respect to this topic, focusing on the need for a well-developed annotation scheme. The desiderata for such a specification language are defined along with what representational mechanisms are required for the specification to be successful. We then review several spatial information annotation schemes, focusing on the latest version of the ISO-Space specification. Finally, we discuss where ISO-Space still falls short and propose some ways that the specification could be enriched.