Abstract
Progress in automatic control, particularly in optimizing and self‐adaptive controls, is used to establish models in order to connect three horizons: knowledge, human activities and automatic controls. The guiding operation in all these horizons is recognished as the evolution. In the horizon of automatic controls, the evolutions can be said closed, since they are completely defined. In the horizon of human activities, they can be said open, because their over all direction depends both on the constituents and on the partial accomplishment. In the horizon of the acquisition of the knowledge, there appear to be both evolutions in closed and open form; the problem is therefore to know whether these evolutions are teleological or homeostatic. There does not seem to be a ready answer; it could however probably be accelerated by better knowing the concept of matching the analogy