Computing with cognitive spatial frames of reference in GIS

Abstract

In everyday communication, people effortlessly translate between spatial cognitive perspectives. For example, a guide leading a group on a hiking trail may need to translate the instruction "turn west at the church building", read from a trail description, into "turn right in front of the church". While the different types of cognitive reference frames and their relevance for different language cultures have been studied in considerable depth, we still lack operations that can be used to transform vague geometric locations from one frame to another. In this article, we suggest that one reason for this is that cognitive reference frames require transformations different from the ones used in current Geographic Information Systems (GIS). These transformations need to take into account vague translations, rotations, and scalings. To approximate the location denoted by a natural language expression like "in front of the church", we consider vague transformation models in terms of neural fields. We introduce a simple and tractable method of computing transformations using fuzzy vector spaces, and show how it can be used to describe a spatial scene in a geographic map from the perspective of six well-known cognitive frames of reference.