The problem identification-design-build-evaluate-theorize structure of design science research has been proposed as an approach to creating knowledge in information systems and in broader organizational and social domains. Although the approach has merit, the philosophical foundations of two specific components warrant attention. First, the grounding of design theory on potentially incommensurate kernel theories may produce incoherent design theory. In addition, design theory has no strong logical connection to kernel theories, and so cannot be used to test or validate the contributing kernel theories. Second, the philosophical grounding of evaluation may inadvertently shift from functionally based measures of utility and efficiency, to evaluation based on the pragmatic fulfillment of multidimensional human actions as people encounter information systems, resulting in evaluation errors. Although design and evaluation from a single paradigm is not desirable, sufficient, or representative of design science research, multi-paradigm grounding of design and evaluation must be realized and used consciously by the research community if the design science approach is to remain a legitimate approach to knowledge creation.