Durlach et al. (2000) outlined a research program for the use of virtual environments (VEs) to train spatial abilities, and have invited comments. Our perspective is that some of the problems facing researchers in VE navigation, and also perceptual-motor skills, stem from the possibility that psychological processes that underlie navigation and motor skills in real and in virtual environments may be somewhat different. VE users may then have to make use of different strategies and abilities to those that they would normally use to manually manipulate and to navigate around the world. We suggest that such differences may arise because, when performing VE motor skills and navigation, the subject is effectively disembodied.