[From Bruce Gregory (990204.1735 EST)]
Since Martin has opened the door to heterodox thought, I offer the
following speculation. The role of awareness is not well-defined in
HPCT--the system works in the same way whether or not awareness is
"present". For some time I have speculated that awareness goes to where
the error is greatest in the system, which is consistent with Bill's
conjecture that awareness is associated with reorganization. I no longer
think that my conjecture is tenable. One reason is the existence of
long-lasting errors associated with goals that cannot be realized at
once. It now seems to me much more likely that awareness goes to where
it is most needed to exercise control.
You pull up behind a car at a stop-sign. You want to avoid hitting the
car in front of your and to cross the intersection with a minimum of
delay. If you keep your attention on the car in front of you, you will
be sure to accomplish the first goal. But as long as your car is
stopped, hitting the car in front of you is not a potential problem--it
does not require your attention to exercise control. Not being hit by a
car on the cross street, however, does require your attention. If you
are not careful to you will allow your attention to focus only on this
goal--when you see the side road is clear you may then drive into the
car in front of you because your attention is on the side street not the
car in front of you.
How might this story be modeled? I can imagine one ECU controlling for
distance between you and the car ahead with a second ECU controlling for
avoiding cars on the cross street. A higher order system might alter the
gains of these two systems to stabilize its perception of "driving
safely". Or the higher order system might alter the reference levels for
the two lower level systems. I am not sure that this model captures what
is happening, however. Rather it seems as if what is being shifted is
awareness. This shift of awareness is required for no other reason that
we can't "pay attention" to the car in front and possible cars on the
cross street at the same time. That is perceptual signals are only
present in one control loop or the other, but not in both. It depends on
where you are looking--where your attention is.
Rick, can you give me a little help here?
Bruce Gregory