Three experiments explored how participants solved a very open-ended generative problem-solving task. Previous research has shown that when participants are shown examples, novel creations will tend to conform to features shared across those examples (Smith, Ward, & Schumacher, 1993). We made the shared features of the examples conceptually related to one another. We found that when the features were related to the concept of hostility, participants' creations contained hostile features that were not part of any of the examples. These results suggest that participants will design novel entities to be consistent with emergent properties of examples shown to them. We also found that a mild hostility prime from unscrambling sentences had a similar conceptual effect. Together, the two effects suggest that conceptual priming of generative cognitive tasks will influence the cognitive aspects of the creative process.