Constraints on priming in spatial memory: naturally learned versus experimentally learned environments.

In four experiments, we explored constraints on priming in spatial memory. In Experiments 1 and 2, subjects who were familiar with the locations of buildings on the Vanderbilt campus participated in a recognition test. The subjects' task was to decide whether or not named buildings were on the campus. Foils in this recognition test were realistic but fictional names of buildings. In principle, the subjects could have performed this task without using spatial knowledge; in fact, they must not have used spatial knowledge, because there was no evidence of priming in recognition as a function of the spatial relations between buildings on the campus. This result differs from those obtained in earlier experiments that have examined memory of spatial layouts learned in laboratory settings. In Experiment 3, the fictional foils were replaced by names of buildings in an area of the campus separated geographically from the main campus. Evidently, this change induced subjects to retrieve spatial knowledge, because the spatial priming effect materialized. A fourth experiment replicated the above findings in a single experiment and demonstrated that spatial priming could be obtained when the configuration of buildings was learned experimentally. These results are explained by appealing to the "decontextualization" that takes place in memory over time.