Classification Structures for Cognitive Maps

The ability to create and manipulate meaningful data structures of cognitive spaces remains a problem for designers of geographic information systems. Methods to represent the inherent hierarchical structure in cognitive spaces are discussed. Several alternative scaling techniques for developing hierarchical and overlapping representations, including ordered trees, ultrametric trees, and semi-lattices, are presented and discussed. To demonstrate the differences among these three representation schemes, each of three techniques is applied to two small datasets collected on the recall of capitals or countries in Europe. The methods discussed here were chosen to illustrate the limitations of a strict, hierarchical representation and because they have been used in the past to model cognitive spaces.