Multiplicity of Mental Spaces

When asked the direction between Phildadelphia and Rome, most people err. They say that Philadelphia is north of Rome when in fact, it is south of Rome. This cannot be dismissed as the weather, because when asked the direction between Boston and Rio, a majority of people erroneously say that Boston is east of Rio. Nor are these errors a simple consequence of randomness, nor of ignorance of geography. Rather, they are systematic and predictable outcomes of the way spatial information is organized in the mind. Contrast these errors with the precise artistry of violin playing or basketball or surgery, or wending one's way through a crowd. How is it that one set of spatial behaviors is predictably erroneous and the other predictably precise? That one set of spatial behaviors appears blunt and irrational and the other delicate and tuned? How, then, to reconcile spatial behaviors that are finely-tuned and precise with those that are clumsy and erroneous? A closer look at the finely-tuned, precise ones reveals that these behaviors are repeated, indeed practiced, in structured environments that provide feedback and that are replete with cues that guide and support behavior. Sheer repetition in supportive environments provide these perceptual-motor performances the benefits of selection, by evolution or by learning. Behavior that yields good outcomes is selected and repeated; not so behavior whose outcomes are not favorable. Rich environments provide cues for the behaviors more likely to yield favorable outcomes. In short, through practice with feedback, actions become precisely tuned. By constrast, spatial behaviors that are infreqent and based on hypothetical rather than experienced environments, notably judgements, do not enjoy the benefits of either repetition with feedback or of supportive environmental cues. Instead of being situated in real environments, they are mediated by cognitive representations of environments that are local, ad hoc, and schematic. Although the cognitive structures constructed to represent environments and to enable judgements on them are based on qualities of the perceptual world, the fact that they are schematic introduces systematic error and the fact that they are ad hoc and local means there is no guarantee of global coherence, introducing the possibility of inconsistency. Ad hoc, local, schematic cognitive structures are invoked when there are no means, either by the tools of the mind or by the tools of the world, of producing complete representations on which appropriate and accurate calculations can be performed. Given the constraints …