Concepts generated by engineering graduate students who received stimulus verbs that were antonyms of the core functions of a problem generated were judged to be more novel than students who receive stimulus verbs that were synonyms of the core functions of the problem. This effect was robust for two out of four problems studied.