Intelligence as a Border Activity Between the Modelled and the Unmodelled

Abstract The common conception of intelligence in terms of information processing has its origin in cybernetics and information technology. Its import into cognitive science and the humanities not only generates theoretical problems, but also constitutes the basis of methods and policies that have adverse impacts on intelligent agents. In order to demonstrate why this technological conception of intelligence is not suitable for addressing the intelligence of living agents, and why natural intelligence is artificially not imitable, first, the basic notions, formalisms, and assumptions of the technological context are presented. Next, the cybernetic feedback scheme is modified by dropping the technological assumptions, and a more sophisticated closed-loop scheme is developed that is better suited for representing cognitive processes in living agents. With reference to the thus developed scheme, the capacity of achieving ontological expansion, via what will be called “internal restructuring,” is proposed as the criterion for intelligence.