Matthew Hutson, English 111, Brown University (1998)
For those unfamiliar with the idea of feedback, think of all those times you've heard microphones go batty and deafen entire crowds with heinous screeches; this is an example of feedback. The point of a microphone is to pick up the sound coming from your mouth and send it to a speaker, which amplifies it and broadcasts it (a fraction of a second after you make the original sound.) When a microphone is held too near a speaker, the sound coming from the speaker will be fed back into the microphone, which will then be amplified by the speaker and fed back into the microphone again, and so begins a feedback process that quickly builds up to an intense, abstract wail of noise. This is an example of positive feedback. Another form of positive feedback actually dampens the process of amplification, so that noise (or other forms of information or energy) will eventually (or quickly) die down to nothing.
Differing in a misleading way from the latter form of positive feedback is negative feedback. This is a process that tends to counteract the change in a system, rather than increase it. Thermostats illustrate this concept; one sets a thermostat in a house to 68 degrees, and the thermostat reacts to deviations from this desired temperature. If the house goes up to 70 degrees, the thermostat will measure this and turn down the heater, and if the house goes down to 66 degrees, the thermostat will tell the heater to work harder. Negative feedback loops are especially important in biological systems; balancing on a tightrope, for instance, requires one continuously to shift one's balance away from the direction of falling in order to stay alive. These systems often require extreme delicacy and sensitivity for success.
Digital feedback systems exist too, but they feel restricted, unnatural. In order to understand their lack of empathic subtlety, one can imagine a simple string of numbers, ones and zeros: 1 0 0 1 1 . . . Now if we apply a transformation to this data such as multiplying it by 9/10 or 11/10, and analog system would give us .9 0 0 .9 .9 and then .81 0 0 .81 .81 or 1.1 0 0 1.1 1.1 and then 1.21 0 0 1.21 1.21, respectively. A digital system would still adhere to the discrete 1 0 0 1 1 in all cases. A living organism operating according to a digital feedback system would not last long within our continuous, analog environment.