Visualizing Egocentric Path Descriptions: A Computational Model.

摘要

The ability to visualize spatial information from verbal descriptions is an important component of human cognition. A common example is generating a mental picture of driving directions. Such directions can be given either from an external viewpoint, as if viewing a map (exocentric description, e. g. Left is always West), or from the point of view of a traveler moving along the path (egocentric description, e.g. take a right, go 1 block, then turn left ). Directions for driving imply a horizontal-plane, two-dimensional mental image, but one can also describe paths through 3D space. We have studied the capacity of people to visualize complex 2D and 3D paths using the Path Visualization (PV) task, which provides an objective measure of visualization accuracy. We developed an ACTR model of visualization capacity for exocentrically described paths (Lyon, Gunzelmann & Gluck, 2008). According to this model, the capacity to visualize an exocentric path description is limited primarily by decay and spatial interference in an exocentric-viewpoint image constructed in visuospatial working memory.