ABSTRACT
Controlled systems of the type that are described in cybernetic models are indeed open to flows of energy but comparatively closed to information. Learning holds a central place in W. Ross Ashby's work and in cybernetic theory generally. Contrary to what happens in living organisms, self-production is only "phenomenal" in the case of cybernetic artifacts. Cybernetic complexity reveals its Janus-like twin faces. It is equally misleading to extoll the liberating potentialities of cybernetics or to condemn it for its coercive implications, if both perspectives are not kept in balance. In the aforementioned classification proposed by J. Attali, "cybernetic information" refers only to the transmission of signals; however, his next level, namely semantic discourse is just as relevant to cybernetics, and casts a new light on its bias. Cybernetics has been criticized also for its inability to conceive regulation as something else than a reaction to "errors."
