[From Bill Powers (960214.0100 MST)]
Martin Taylor (960213 17:30) --
In answering Rick Marken you said
If you have an observation of a variable that is clearly being
disturbed by some unobservable cause, but which nevertheless
persists in returning to or close to its original position, how can
you tell whether the variable is or is related to the CEV of some
control system?
If the cause of the disturbance is unobservable, you can't do the Test.
I think we have to find a way of speaking consistently about "the
disturbance" (the disturbing variable or disturbing influence) and the
_effect_ of the disturbance. How about using the term "perturbation,"
which I think always refers to an actual effect on a variable?
When a disturbing influence is applied to a CEV, it is expected (when
there is no control) to produce a calculable perturbation of the CEV. If
the actual perturbation observed is less than the expected one, there is
probably control. Note that in this way of putting it, there need be no
observable perturbation -- control may be so good that close observation
would be needed to detect that there is any perturbation at all. If the
gain of the control system is high and the disturbance bandwidth is well
within the control bandwidth (as it should be to do the Test properly),
the observed perturbation can be comparable to the noise level of the
observation.
Also, I once again remind everyone that the Test does not consist only
of the application of disturbances and observations of perturbations.
The lack of an effect of an applied disturbance could be due to an
incomplete understanding of the physical situation, or some other system
beside the one you have in mind might be doing the controlling. To
complete the Test, you have to verify that the appropriate input and
output connections are present, so you know which system is doing the
controlling and what the input and output pathways are. The
identification of the input pathway is done by interrupting it and
demonstrating that control is lost (under Hans' model, this might take
some time).
···
-----------------------------------------------------------------------
Remi Cote (960213.1600 EST) --
I wanted to ask myself publicly, can we imagined an organism
becoming madder and madder because he has to much control? Of
course I learned that this list is not the right place for that
kind of questionning.
It's not too much control that drives people mad; it's trying to control
what is basically not controllable.
A person who is obsessed with controlling everything is trying to do
more than is possible, and is probably failing to control many things as
a result. A particular problem arises when a person tries to control
other people, because if they are also controlling the same aspect of
the world that our controller wants to control, the result will be
conflict. It's also possible that when you try to control too much --
the behavior of the Mississippi River, for example -- you don't
understand the true problem, and in your attempts to control one thing
you cause worse problems somewhere else.
Also, the term "self-control" really means conflict with oneself, as the
term is normally used. If there is no conflict, one simply carries out
what is intended, with no internal opposition. In fact, most languages
have words for unconflicted control -- in French one word is "faire", in
German "machen" or "tun", and in English "do" or "make." Unconflicted
control does not feel like what most people think of as controlling. If
you want to write your name, you simply do it. If you want breakfast,
you make it. The fact that these are control processes does not make
itself obvious. True control without conflict feels effortless.
Another aspect of the problem is the sensitivity one brings to a control
process -- sensitivity in the sense of how small an error one still
tries to correct. As Hans Blom has pointed out, perfect control is not
only impossible, it is generally undesirable. A person who tries to keep
a house absolutely clean and orderly will expend large amounts of time
and energy going over and over the house, removing every tiny speck of
dust as it falls and adjusting all the furniture a millimeter this way,
a micrometer that way. What should be a finite and satisfying task
becomes an endless burden, leaving no time to do anything else, even
more important things.
Control sensitivity has two major determinants: the acuity of
perception, and the output sensitivity which determines how much action
will be generated by a given amount of error. If the house-proud person
goes over a clean house with a magnifying glass, there will be more dust
seen that calls for removal. And at a given magnification, the person
might remove obvious specks of dust, or insist (the word Martin Taylor
has suggested) on removing every last visible particle.
Although people are, by their very nature, control systems, their
conscious management of their own lives (at the higher levels of
organization) is not very skilled, simply because the phenomenon of
control has not been understood. People have great difficulty getting
along with each other because they don't really grasp the fact that
other people are control systems, too, with their own goals and agendas.
Many human institutions, from the family to the government, are
organized around controlling other people. When attempts to control
others run into difficulties, as they always do, the problem is seen as
one of not having enough control, and the efforts to control are
increased, which only makes the problem worse. This is why we keep
having to build more jails, and "crack down" with more energetic law
enforcement, and increase the severity of punishments. Such actions
increase the level of conflict, and produce increased resistance that
calls for even stronger actions.
So you see, Boy Remi, that I do not believe it is control that drives
people mad: it is ignorance about control that causes the problems.
----------------------------------------------------------------------
Best to all,
Bill P.