Subject: Feedback Loops

Posted by John Sterman on 2/8/2006
In Reply To:Feedback Loops Posted by Phil Minnaar on 2/7/2006

Message:

let me take a stab at the loops in Phil's examples.

Lightning: The molecules heated by the lightning transfer energy to others by radiating high-energy photons and by collision with other molecules in the air (and perhaps through other modes such as ionization). Each of these dissipation modes forms a negative feedback of the form: Energy of molecule ->+ energy transfer to environment -> - energy of molecule. The high-energy molecules transfer energy to others until (on average) their energy falls to the same level as their surroundings (basic negative feedback underlying thermodynamics).

Drawing loops: Every decision is involved in at least one feedback loop, but not every causal pathway may form a loop (at least over the relevant time horizon). The existence of 'dead ends' in one's causal maps does not indicate that there are not loops, just that those pathways don't appear to feed back. Note the loop implicit in the process as well: on finding a dead end, one may reconsider the path, the data supporting the links in the path, or start to follow a different path, some of which will eventually feed back.

Top Management: Let's stipulate for the moment that top management is completely disconnected from the operations of a department in an organization. The decisions made by the people in that department are still governed by feedback. You don't specify particular decisions, but these typically would include how people allocate their time among competing activities, production (of widgets or memos or whatever), etc. All of these are involved in multiple
loops: the more work there is to do, the more time people will put into it, reducing the work remaining, etc.

John Sterman