There is, in the ®eld of education in the English-speaking world, an enduring faith in the capacity of social engineering. Witness the`What works' syndromeÐa quest for assured knowledge to back up pedagogical practice and curriculum-making. The metaphor is that of a machineÐa complex one grantedÐbut a machine nevertheless. It is the sort of machine that Norbert Weiner imagined the body to be, or the way organisms interacted, when he developed his cybernetic model. Ironically, a key to the way his system worked was negative, corrective feedbackÐnot what works but what doesn't. Positive feedback tended to set systems out of control, he said. We can see the e ects of positive feedback in the case of computers which are purchased in great numbers for schools, although little is known about what they contribute to educationÐand least of all what teachers actually do with them.Research in education needs to ®nd out what teachers think of reformÐ to ask those who have intimate knowledge of what happens when grand schemes are launched. Such a need to consult those who do the work can be seen most dramatically in the case of nurses, who, in Canada at least, are now being recognized as sources of important information for purposes of assessing how hospitals work. They, like teachers, can and will in the future provide the negative feedback most often ignored and perilously so. However, does a systemic approach to reform in education appreciate the value of negative feedback. I think not.The systemic approach to reform recognizes that schools are part of a complex system of expectations and that reform plans must recognize the diverse interests at play in any reform. The idea is that when all the parts are pulling together reform will happen. The assumption is that the purposes for schools can be established, and that the technical means identi®ed to achieve those purposes are known. It is a matter of everyone being`on-side'. To me, this sounds like the roll-call and the rallying of j. curriculum studies, 2002, vol. 34, no. 2, 129±137 John Olson is professor emeritus in the