Item

Abstract

known to make extensive use of geographical slant for communication about spatial layout (Brown and Levinson, Journal of Linguistic Anthropology 3: 46-74, 1993). We have
investigated the role of geographic slant in a simple spatial memory task. The experimental environment is a modified version of the Hexatown virtual environment described by Gillner and Mallot (Journal of Cognitive Neuroscience 10: 445-463, 1998). This environment
is a hexagonal grid of streets with landmarks placed in each angle between two streets. We used a version with 8 places and no loops. The whole environment could be
slanted by an angle of 4 degrees. Three slant conditions were used: “flat”: no slant; “slanted NW”: slant direction 30 degrees north west with respect to some arbitrarily chosen “north”; slanted NE: slant direction 30 degrees north east. Subjects could interact with the virtual environment by pedaling with force-feedback on a bicycle simulator (translation) or by hitting buttons (discrete rotations in 60 degree steps). The environment was simulated using a Silicon Graphics ONYX2 computer. Images were projected on a 180 degree screen. For details of the setup, see van Veen et al. (Future Generation Computer
Systems 14: 231-242, 1998).
Subjects explored the environment by searching 15 routes to various goals presented to them as pictures. After learning, spatial memory was accessed by a pointing task: In an across subjects design, 3 groups of 6 subjects were asked to point from various positions to the learnt goals (19 pointings per subject). Overall performance is rather good, with a mean angular error of -5.9 degrees (plus/
minus 53 degrees mean angular deviation) in the “flat” condition. Performance was significantly better in both the “slanted NW” condition (circular F-test, p<0.00001) and the
“slanted NE” condition (p<0.04). There is also a significant difference between the two slanted conditions (p < 0.01).
The results show that subjects are able to point to currently invisible targets in virtual environments. Adding a geographic slant improves this performance. We conclude that geographical slant plays a role either in the construction of a spatial memory, or in its readout, or in both.