Knowledge of space is critical to attitudes toward, decision making about, and behavior within places. A person's environmental knowledge depends upon how knowledge is acquired and mentally processed. Past studies of environmental knowledge acquisition have focused primarily on that which results from behavior in the environment. Research has led to a developmental theory that views knowledge acquisition as a gradual process that begins with selective, fragmentary information about a place, to which the individual information adds over time until an integrated cognitive representation is achieved. The present study was designed to determine the applicability of this developmental theory of environmental knowledge acquisition to learning from maps. I also consider issues of hierarchical spatial knowledge organization, the extent to which cognitive "maps" are images, propositions, or both, and the systematic biases inherent in knowledge obtained from maps. This paper describes an experiment in which map presentation strategies incorporating segmentation are compared to a control case. Two of these segmentation strategies are designed to facilitate a developmental sequence for knowledge acquisition, one focused on landmarks and the other on routes. The third segmentation strategy emphasizes development of a regionalized hierarchical knowledge structure. The developmentally-based segmentation strategy emphasizing routes makes knowledge acquisition easier, but the control case in which the map is not segmented results in faster access of distance and direction information on a subsequent test. No difference in accuracy of distance and direction estimates occurred among groups. Overall findings support a dual coding theory for spatial knowledge in which both analog (image) and propositional information is included in cognitive maps. Distance and direction tasks seem to be solved using imagery except when subjects judged to have low visualization ability use a map learning strategy that facilitates propositional encoding-the route-segmentation strategy. Evidence suggests that differences in map versus environment-derived cognitive representations identified by other authors are not simply due to a difference between viewing maps holistically and environments in segments, but to a combination of segmentation with a focus on routes that leads to procedural encoding.
