Feedback

Control engineering, 1920s; generalised by cybernetics in the 1940s

The return of a system's output to its input, so that what a system has done shapes what it does next. Negative feedback stabilises; positive feedback amplifies.

Feedback is the basic mechanism of cybernetics. A system acts, the effects of its action are sensed, and that information shapes its next action. The causal chain closes on itself into a loop.

Negative and positive feedback

In negative feedback, the system acts to reduce the difference between where it is and where it should be. A thermostat switches the heating on when the room is cooler than its set point and off when it is warmer. Body temperature, blood sugar and a driver steering along a lane work the same way. Negative feedback produces stability and goal-seeking behaviour.

In positive feedback, a change produces more of the same change. Microphones squeal when the sound from a speaker is picked up and amplified again; panics, bank runs, viral content and melting ice that exposes darker, heat-absorbing water all follow this pattern. Positive feedback produces growth, escalation and sudden shifts between states.

Real systems combine both. Many problems come from delays in a loop, which cause overshooting and oscillation, or from positive loops that grow until something breaks.

In the history of cybernetics

Engineers had used feedback in governors, amplifiers and servomechanisms long before the 1940s (James Clerk Maxwell analysed governors in 1868, and Harold Black invented the negative feedback amplifier in 1927). The cybernetic step, taken in “Behavior, Purpose and Teleology” (1943), was to argue that purposeful behaviour in animals and people is also explained by negative feedback. That turned feedback from an engineering technique into a general principle connecting machines, organisms and societies.

Further reading

  • Otto Mayr, The Origins of Feedback Control (1970)
  • Donella Meadows, Thinking in Systems (2008)

Mentioned in