This paper presents a methodology for the design of intelligent learning environments. We recognise that in the educational technology field, theory development and system-design should be integrated and rely on an iterative process that addresses: (a) the difficulty to elicit precise, concise, and operationalized knowledge from "experts" and (b) the crucial differences between the communication modalities that experts can relate to, and those that are available to a computer-based system. Inspired by the well-known "wizard-of-Oz" methodology we discuss the need for characterizing and carefully controlling the range of its possible variations. We refer to our approach as "tapering" of the communication capacity of carefully engineered didactical situations and present its application, and a case study from our work with an exploratory environment. We then discuss the generality of the methodology and pragmatic constraints which can be useful in similar research. (Contains 7 figures.)