In this article, Ralf St. Clair makes the argument that induction--the process of applying research findings from one setting to another--is logically unsupported, irrespective of method or methodology, due to the existence of superunknowns. Superunknowns are defined as factors that cannot be anticipated, not because of instrumentation defects, but because of their nature. On this basis, St. Clair asserts that we cannot increase the credibility of educational research by trying ever more strenuously to create general laws. Instead, St. Clair argues for educational research to be viewed as a means to generate empirical heuristics for thought and inquiry, and for wider recognition of the central--and essential--role of human judgment, as exercised by practitioners and researchers, in the research endeavor. (Contains 5 notes.)