The importance of iteration in creative conceptual combination

Theories of creative conceptual combination hypothesize that, to generate highly creative concepts, one should attempt to combine source concepts that are very different from each other. While lab studies show a robust link between far combinations and increased novelty of concepts, empirical evidence that far combinations lead to more creative concepts (i.e., both more novel and of higher quality) is mixed. Drawing on models of the creative process, we frame conceptual combination as a divergent process, and hypothesize that iteration is necessary to convert far combinations into creative concepts. We trace conceptual genealogies of many hundreds of concepts proposed for a dozen different problems on a large-scale Web-based innovation platform, and model the effects of combination distance on creative outcomes of concepts. The results are consistent with our predictions: (1) direct effects of far combinations have a mean zero effect, and (2) indirect effects of far combinations (i.e., building on concepts that themselves build on far combinations) have more consistently positive effects. This pattern of effects is robust across problems on the platform. These findings lend clarity to theories of creative conceptual combination, and highlight the importance of iteration for generating creative concepts.