Search results

The purpose of this paper is to present a computational approach and a toolset to infer spatial displacements as they occur in route narrative documents and report on first experiments done to produce computer‐aided learning (CAL) applications and instructional design editors that exploit the inferred georeferenced itineraries.

Identification of the structure of natural language expressions to indicate spatial displacement, as they occur in route narrative documents. Use of geographic information systems technology to treat the semantics of spatial displacements identified in route narrative documents.

It was found that French route narratives can be linguistically analyzed with a relatively simple structure (two computational models). Transducers are used to successfully analyze displacements verbs and to attach some semantics to paragraphs of route narratives.

The same structure valid for French texts can be found in English texts as well. Further work could lead to fundamental questions about whether itineraries are linguistic universals.

The toolset is fully functional and can be tested on the World Wide Web for information retrieval purposes. Currently it is used for CAL applications to tutor learners in the process of reading route narratives documents.

The paper demonstrates that the featured computational models and toolset are mature enough to be embedded in end‐user applications.