Abstract: This paper puts forward the view that the life-cycle model of system development has been allowed to make the transition from a useful intellectual tool for discussing specific aspects of the process to a definitive statement of what the process actually "is". We argue that this hinders the discussion of Requirements Engineering in fundamental ways, and that challenges to the supremacy of this model are needed to open up effective debate. This is illustrated by the introduction of a model based on the construction of theories, which shifts the emphasis from a technocentric view of the process, where requirements are encoded in initial forms of a system, to one focused on the human understanding of situated systems. Thus requirements become the forces and constraints which influence the designer. We conclude by presenting samples of areas that can be discussed more effectively from this viewpoint.