International Encyclopedia of Systems and Cybernetics

The achievement of a satisfying global vision of a system involves a difficult methodological problem: Synthesis. In spite of its importance it did not worry too many systemists up till now, except J.L.LE MOIGNE ("Theorie du système général", 1990a) and R. RODRIGUEZ DELGADO (1993b) This last author distinguishes the following different ways toward synthesis:

"-Structural synthesis: Once structures have been analyzed and well known they should allow for a global and renewed representation of the system.

"-Functional synthesis: the same structures may lead to various functional distributions, which should be optimized. Synthesis and generalization methods allow for the global study of functions and the shaping of functional models which will permit to select some convenient ones.

"-Relational synthesis: synthesis and generalization methods also may be used for the shaping of models of the systems internal and external relations and of criteria allowing for the best adapted choices respect to the problem or the situation as it is.

"-Historical synthesis : historical synthesis studies allow for a global view of the system development in time and for predictions which would not be possible by other techniques.

"-Interdisciplinary synthesis: which takes place when two or more disciplines formerly isolated become united as for example: Physico-Chemistry, Biochemistry, Sociocybernetics, or Ecology.

"-Prospective synthesis: by use of feed-forward techniques it becomes possible to better short-term predictability and to establish probable medium and long term scenarios.

"Synthesis techniques vary widely according to the class of systems to which they are applied" (1993b, p.293-4).

One may also think that the various synthesis methodologies are in every case depending on specific systemic concepts.

For example:

- Structural and functional synthesis should be grounded on a common taxonomy of subsystems (MILLER, 1978) and on the concept of synchronic-diachronic structure.

- Relational synthesis should use interconnections graphs and anew the taxonomy of subsystems, as well as the concept of interface.

- Historical synthesis does not seem yet to have found its own systemic and cybernetical models. The concept of limit cycle and those of morphogenesis, steady state and ageing would probably be useful.

- Interdisciplinary synthesis involves finding new objects for research which could be modelized as complex systems as for example cyclical catalytic reactions, synaptic connections in neural networks, autopoietic organizations or biotopes.

- Prospective synthesis depends upon skilled use of the concepts of bifurcation, catastrophe, and combined cyclical rhythms, among others.