A basic epistemological conflict is found to exist between modern and non-modern practitioners of sustainable development. These categories distinguish the ways professionals interpret or frame reality. The hypothesis developed is that this inconsistency, at least partially, explains the limited success that energy-efficiency research has realized in the prediction and control of climate change catalysed by the built environment. An analysis employs both historical and empirical methods to understand how the North American air-conditioning industry has framed, and subsequently regulated, the inseparable problems of human comfort and energy consumption. Historically, the dominant framework long-inhabited by moderns has constructed a unit-efficiency model of evaluation that is concerned with universal standardization and normal design. In the empirical analysis of the selected case, an emergent framework inhabited by non-moderns constructed a unit-efficacy model of evaluation concerned with local implementation and post-normal design. The two models came into conflict when designers applied code-required energy models and financing formulae based on unit-efficiency assumptions to a case of sustainable, affordable housing. The analysis concludes with seven findings designed to move building energy research and practice beyond the current epistemological divide.