International Encyclopedia of Systems and Cybernetics

A system able to absorb and emit energy and information, but not matter.

This definition is generally accepted.

Notwithstanding, the concept of closed system is one of the most semantically dubious in systemics.

To begin with, it should clearly be distinguished from the isolated system concept, with which it has been sometimes confused. Indeed, the isolated system does not absorb or emit anything… and, as a result, cannot be perceived by any observer.

It has been also been written by some authors that the closed system exchanges only energy with its environment, excluding matter and information. It is quite difficult to imagine a concrete system that would offer such a curious characteristic.

Organizationally (or operationally) closed systems are also frequently considered closed in this sense, being then defined as a set of relations between components, such that states transitions of the system are state determined (H. MATURANA, 1975). However, concrete systems that maintain their internal organization are able to do that even when they exchange matter and, or information with their environment. Only that some exchanges are admissible and others are not. Obviously "closure" does not mean insulation or isolation.

It could be argued that any satisfactory model of a concrete system must necessarily and simultaneously be open and closed, according to the functional requirements of the system. Thus we should speak only about closed (or open) systems specifying the inputs and outputs they admit or not, as well of the precise moments these are admitted or rejected.

N. GEORGESCU ROEGEN formulates the following basic law: "No closed system can perform mechanical work at a constant rate indefinitely" (1982, p.318). He notes that this law "… is the twin of the second law of thermodynamics, which prohibits the endless derivation of mechanical work at a constant rate from a finite amount of free energy" (Ibid, p.321).

Another problem with the closed system model has been emphazised by G. BROEKSTRA (after G. SAGE): "The closed system metaphor was that of the organization as a machine or clockwork" (1992, p.1025). The presence of this metaphor in the minds of many business executives has undoubtedly provoked the demise of numerous enterprises. It has also be instrumental to the awful bureaucratization of many organizations at the administrative and governmental level. Indeed, when the clockwork is constructed, it is not supposed to be modified outside of its own basic rules and it becomes too perfectly "autopoietic", loosing any adaptive capacity".

As noted by T. BAUMGARTNER et al, "… in general, the assumptions (or theory) underlying closed systems models and methods contradict the assumptions required (in practice) in a research situation involving open systems. Rarely, even under experimental conditions, do social scientists achieve the degree of closure (or control) which closed system models and methods require" (1976, p.39).

There is, furthermore, a time frame problem: "Indeed, social science data and analyses are often based on a time span of observations so short that the "open" nature of the system is not detectable, and relationships between and among variables appear stable" (Ibid).

And this is not only true in human systems, but also in complex systems in general.