Hidden Structure: Indirect Measurement of Relational Representation

Structured mental representations underlie much of human cognitive ability. However, research has repeatedly found that people are generally quite poor at spontaneously applying structure acquired from one kind of situation to a dissimilar domain. This would seem to present a serious impediment to learning abstract structures experientially. The current study employs a novel approach designed to be more sensitive to the relational learning that is taking place. By using implicit measures of processing fluency, we demonstrate that relational commonalities—even between dissimilar domains—can indeed exert a measurable, if subtle, influence on subsequent processing. Despite having no explicit awareness of the structural commonalities, participants in our study rated a scientific passage to be better understood, better written, and more interesting when it was preceded by an (overtly dissimilar) analogous passage. This finding has important implications for the acquisition of complex knowledge structures.