Selective Attention and the Acquisition of Spatial Semantics

Abstract

The acquisition of the semeintics of natural language spatial terms is considered within the cognitive framework introduced by (Langacker, 1987), and the computational framework of the Berkeley Lo project (Feldman et. al., 1990). We describe a computational model which incorporates selective attention mechanisms to facilitate the identification of significant objects within the visual field, and their consequent binding to linguistic relationeil identifiers (for example, the trajector and landmark) according to the conventions of the input Uinguage. In contrast to previous work in this area, the approach allows extension of the system to more sophisticated (potentiiilly cluttered and feature-laden) input scenes 8ind referential linguistic phenomena, without a major redesign of the system. The application of the model to lexemes describing static concepts such as the English above, below and in is discussed, as are extensions to dynamic concepts.