Subject: Feedback Loops

Posted by John Gunkler on 2/9/2006
In Reply To:Feedback Loops Posted by Prof. Dr. Niall Palfreyman on 2/8/2006

Message:

As Barry Richmond used to say: Grass causes cows, which cause more grass.
That is, grass causes (through its ingestion) cows to grow, which cause (through their manure) grass to grow.

From this example, I think two things are obvious:

1. We use "causes" in a funny way, but one which is probably more consistent with reality than the one-way definitions we were taught in school. (Of course, in Barry's example, we're also using a shorthand -- we're actually saying that the ingestion and digestion of grass contributes to the growth of the cows who eat it, etc.)

2. A feedback loop does not have to feed back everything in order to be a loop. That is, the cows' manure does not have to be the only source of nutrients for the grass, and yet the manure is still making a feedback loop that fosters more grass growth. Nor does the grass have to supply everything a cow needs to grow in order for there to be a loop. Nor does everything about being a cow have to be contained in the feedback loop.

When you talk about information feedback this is even more obvious -- the feedback "signal" is information about the material that constitutes its source (the stock), not the material itself. So a feedback signal (information flow) that contains information about the size of the stock can control whether more material (material flow) is added to the stock or not.

It is not heat that is transmitted in the feedback loop of a thermostat, it is information about heat that controls whether the furnace comes on or not.


John

P.S. Niall, just because we all do not articulate the same definition does not necessarily mean there is not a fundamental definition for feedback.
There are certainly technical definitions that are well accepted. The challenge we were given, however, was to come up with one that is useful when teaching children or a naïve audience. All science has this same kind of problem -- agreement on technical definitions of concepts, but difficulty in explaining the concepts to lay people. Ordinary language and scientific discourse are different animals, unfortunately. We are very sloppy in our ordinary language, and some of this ambiguity contributes to our delight in using it because it conjures up emotional responses and rich contexts -- which is fun. Science cannot usually afford this kind of communication.

Follow Ups:

Feedback Loops - Bill Ellis 2/9/2006