An investigation into the psychology of spatial and temporal reasoning

The present work investigates the logical abilities of naive reasoners when given relational reasoning problems. Subjects were given one of two versions of the relational reasoning task: a temporal version or a spatial version. Both versions of the task consisted of two passages of text from which the reasoner was asked to draw certain deductively valid inferences. The temporal task required the subjects to determine what temporal relationships must hold among the events mentioned in the text, while the spatial version of the task required the subjects to determine the spatial relationships that must hold among the objects mentioned in the text. A model of how the reasoner adopts those conclusions is proposed and tested in the dissertation.;The subjects were given a booklet which contained either spatial or temporal versions of the problem. It was discovered in both cases that the subjects tended to invalidly linearize the component events/objects when asked to determine the temporal/spatial relationships involved. That is, subjects tended to erroneously conclude that the text implied a linear ordering among the component events/objects. The tendency to linearize, however, did not emerge for all of the problem types presented to the subjects.;The second experiment determined that the result mentioned above is due to an error in reasoning, not to an interpretation error. That is, it was determined that subjects had no tendency to misinterpret the individual sentences in the text. The third experiment tested whether the linearization tendency could be overcome by a change in the instructions that accompanied the task. The change in instructions had no effect on the linearization tendency for either of the problem versions. The final experiment attempted to determine the features of the texts that produced the linearization tendency. It was determined that only a particular text type failed to yield the linearization tendency; further demonstrating the robustness of the effect. Furthermore, it was determined that the direction that the linear ordering tendency could take also depended upon features of the text.