Knowledge-Dependent Models of Spatial Behavior
C. Gustav Lundberg · Geografiska Annaler Series B Human Geography · 1991
The implicit and explicit role of knowledge in models of everyday spatial decision making is examined, and knowledge-based models (KBM:s) are suggested as a means for bridging the methodological and philosophical gap between models of discretionary and constrained spatial behavior. The argument is rooted in the theory of heuristics, the methodology of KB-modeling, and central social science concerns about contextual/structural factors and constraints that affect human behavior. The focus is on psychologically justifiable (process descriptive) rather than normative decision models. Certain qualities make the KBM framework applicable to both discretionary and spatio-temporally and socially constrained behaviors: i) Choices and behaviors are explicitly assumed to be contingent upon the history and the situational characteristics of the actor(s), and ii) KBM:s explicitly deal with practical knowledge, and emphasize the relationship between knowing and doing. Whereas the constraint-based models explicitly work with a variety of physical and cognitive constraints on an agent's behavior, models of discretionary behavior implicitly or explicitly must incorporate constraints stemming from the agent's domain knowledge. From a KBM perspective, normative decision making models put too big demands on the agent's short term memory while ignoring the agent's powerful long term memory. It is argued that KBM:s can solve some of the problems related to choice sequences, internal versus external constraints, context effects, aggregation, and spatial problem solving, and that KB-constructs allow modelers to maintain greater environmental realism than competing models. Also, KBM-methodology could be instrumental in alleviating the central bottle-neck of structuration theory: the concretization and operationalization of the term knowledgeability.