Knowledge of Spatial Relations: The Specification of the Information Used in Making Inferences

Abstract

On being shown the names of two towns geography students determined whether one of the towns was north (say) of the other. Reaction time decreased as the distance between the towns increased. This finding was explained by supposing that subjects attempt to make inferences (about whether one town is north of another) by first accessing stored information that specifies a town's location only crudely; when this is insufficient to make the relevant inference more finely discriminating information is accessed. The explanation was supported in a second experiment: when one town was Scottish and the other English, reaction times were shorter than when both were English. This result was predicted from the assumption that the information first accessed specifies which country a town lies in: when one town is Scottish and the other English, this information can be used to make the correct inference without further specification of location (since all Scottish towns are north of all English towns), whereas when both towns are English further information must be accessed. Control conditions showed the result was not due merely to a greater familiarity with Scottish towns. The explanation was discussed in relation to accounts of reaction times obtained when other characteristics (e.g. size) of named objects are compared.