The homogeneity principle states that the outputs of a system are always directly proportional to the input. In other words, twice as much in results in twice as much out. This principle, however, can’t predict the effect of the previous states of the system on the outputs. Often previous states have a significant effect on present or future conditions. In other words, linearity can’t capture feedback; inputs and outputs come and go without any connection between them.
Systems operating in space and time will inevitably be affected by feedback loops from their environment as every action the system takes will drive some sort of consequence. This consequence will feed back to the future state of the system that in return will elicit a reaction from the system. Therefore, as soon as we position our system into the real world, things unavoidably turn nonlinear. The more feedback loops we capture and include in our system mapping model, the more realistic of a picture we get about reality — and the more nonlinear the world becomes.
— The Systems Thinker – Dynamic Systems: Make Better Decisions and Find Lasting Solutions Using Scientific Analysis by Albert Rutherford