Behavior Profiling, Challenge

[From Rick Marken (950913.1000)]

David Longley (950912) --



The main issues covered include the relative merits of adopting an
entirely behavioral or third person approach to managing inmates rather
than attempting to work to include an inmate's point of view.

I can't imagine a less hospitible place to have dumped this stuff than here
on CSG-L. Your approach to human behavior explicitly rejects taking the
behaving system's point of view into account; control theory, on the other
hand, shows that it is impossible to understand behavior _without_ taking the
behaving systems's point of view of into account. The problems with
your approach (from a control theory perspective) start on page 1 of your
executive summary. If the assumptions about behavior made in this summary are
wrong (and control theory suggests that they are) then (I'm afraid)
everything that follows is wrong too.

You conclusions about behavior ("extensional stance") are apparently based on
logic; the control theory approach to behavior is based on logic (modelling)
AND empirical test. Given the amount of work you have put into your
behavioral profiling system, I think it is unlikely that you are likely
to be convinced by emoirical test that your fundamental assumptions are
wrong. Similarly, given the amount of work some of us have put into testing
the control theory model of behavior, it is unlikely that you will be able to
use logic to convince us that our fundamental assumptions about behavior are
wrong. So I think it's better to call this relationship off before we all get
too emotionally involved;-)

Bill Powers (950913.0825 MDT) --

So now all that remains is for you to implement this plan as a Pascal
program using MAINDIST as the disturbance.

Hans Blom (950913b) --

Don't change the rules in mid-game. We had agreed on an unknown k!

It's unknown to the model.

Now how about getting to it, Hans. x(t) = u(t) + d(t). You write a model-
based control system that generates u(t) and keeps x(t) = r(t).