FEEDBACK (Second order or double)
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(1) |
| ID | ◀ 1257 ▶ |
| Object type | General information, Methodology or model |
A feedback which modifies any primary feedback, thus changing the former response of a regulator to some class of inputs.
When a system's parameter undergoes a variation, the first reaction of the system is processing the information it receives about the change. This generally leads to a first feedback through which the system adapts itself, according to the existing regulating mechanisms that it already possesses, which become activated, as they are.
However, the system may need to modify the regulators themselves in order to better their action, as needed when its responses, based on the regulators, as they are, become unsatisfactory. This process is basic for learning.
W.R. ASHBY explains it in the following way:
- “The basic rule is equivalent to the following formulation:
- “1 — When the essential variables (of the system S) are not all within their normal limits (i.e. when the trial has failed), no state of S is to be equilibrial, for the rule is that S must go to some other state.
- “2 — When the essential variables are all within normal limits then every state of S is to be equilibrial” (1960, p.84). (ASHBY's first point is interestingly close to H. SELYE's concept of “General Adaptation Syndrome”, but is obviously more general).
ASHBY pursues: “… any system that has essential variables with given limits, and that adapts by the process of testing various behaviors by how each affects ultimately the esssential variables, must have a second feedback formally identical (isomorphic) with that described. This deduction holds equally for brains, living or mechanical” (p.85).