In work published in the 1980's Yujro Hayami and I elaborated a theory of institutional innovations in which institutional changes are induced, on the demand side, by changes in relative resource endowments and technical change and, on the supply side, by changes in cultural endowments and advances in social science knowledge. In the mid-1980's I initiated a research program to explore what development economists might learn from research by other social scientists working in the field of development. In this paper I draw on this earlier work, and on related literature to explore the conditions under which interdisciplinary imperialism or interdisciplinary collaboration can be most productive. I argue that when the objective of research is to advance fundamental knowledge in the social sciences imperialism can be highly productive. But where multiple sources of knowledge must be drawn on for policy, mechanism, or system design interdisciplinary collaboration is essential.