AbstractGeneral public discourse on science and technology is based almost exclusively on the linear model. Other ways of viewing the situation are explored by ‘unpacking’ more dimensions, introducing a ‘department store’ model and knowledge domains in cognitive space, and ending up with a neural net model which allows for ‘learning’ to take place much as in the human brain. This simplistic investigation cannot be a guide for policy-makers but might throw light on certain puzzling features of contemporary R&D institutions.