Mapping and Inference in Analogical Problem Solving – As Much as Needed or as Much as Possible?

There exist many computational models of analogy making, which are based on different assumptions about the cognitive processes which are responsible for successful transfer from a base to a target domain. In most models, mapping of the base structure to the target is seen as the core of analogy making. We believe that the way in which this mapping is realized is a crucial aspect by which the different models can be distinguished. More specific, we identified systematic (as much as possible) versus pragmatic (as much as needed) approaches to mapping as critical assumption by which the computational models can be discerned. In case of pragmatic mapping, only such aspects of the base domain are carried over to the target which address the current focus of interest. In contrast, in case of systematic mapping, the greatest relational structure which contains the aligned parts of base and target is carried over. In consequence, in systematic approaches, many aspects of the target problem are inferred at one time and available for future use, while in pragmatic approaches new questions concerning the target domain may require that mapping and inference must be performed again. We conducted an experiment in the domain of physics to investigate which type of mapping is realized by human problem solvers and got significant evidence for systematic mapping.