[From Rick Marken (920324 13:00)]
Fortunately, I was called away to a big, fun departmental lunch break
before I had time to complete my exasperated reply to Beer's very
polite post (920324). Now I come back and see that Bill Powers has
handled the situation for me with his usual aplomb. Thanks again Bill.
There was a part of the Beer reply I was working on that might be of
interest (it was also the only non-exasperated part). Here it is:
Another example of this might be the moth that folds up its
wings when it detects a bat's sonar signal. I personally find it strange
talk about this as a negative feedback system in which the moth's
perception of the bat's sound is controlled. Moth's can hear the
bat's sounds for great distances. Once it begins to fall, this
perception doesn't suddenly go away.
But it does tend to decrease!
Also, if the bat continues to
pursue the moth, it does not take any further evasive action in an
attempt to minimize the deviation of its BAT-ATTACK signal from the
desired reference level of zero.
Apparently, once it's on the ground there is not much it can do but stay still.
Does it curl into a ball on the ground if the sensed sound get's loud enough?
If it does -- and this curling has NO effect on SENSED sound level, then
you are looking at the pure, SR system. This makes me realize that the
exact same input-output function can be feedback control in one situation and
SR in another. It is the existence of the simultaneous R-->S FUNCTION
that makes the system a control system. If there is no R-->S function,
or if it is near zero then for all practical purposes it is an SR relation-
ship. So -- if the moth really does curl into a ball in response to sensed
sound when it's on the ground just as it does when flying; and if the ground
curling has virtually NO EFFECT on the sensory cause of the response, then,
IN THAT CASE, the moth's behavior is, I think, SR.
I just can't believe that the moth actually acts this way; if it curls on
the ground then maybe the bat could can detect its movement-- or,
if the moth actually does respond to sensed sound that way when on the ground,
then I would think that the curling would fool or have some other
influence on the bat so that the bat tended to wander off (and thus,
reduce the sensory input that caused the curling).
But let's assume that it does respond to sound the same on the ground as
in the air -- AND that the curling, in that case, has no effect on sensed
What this would shop is that the SAME SYSTEM can change from an SR (open loop)
to a closed loop control system just as a result of changes in the physics
of its environment. Again, I would imagine this is actually rare. But it
demonstrates that the unusual properties of a control system (such as the
fact that it controls rather reponds to its sensory input) result for the
negative feedback loop in which the system normally exists. If you can
somehow break that loop (by removing the influence of output on input) WITH
NO CHANGE IN THE INTERNAL STRUCTURE OF THE ORGANISM then the organism
becomes an SR device. Actually, I did this in one of my studies where I surrep-
titiously broke the R--S connection from handle to display in a tracking task.
The behavior of the subject became SR -- until they noticed that there was
something fishy going on (higher level perception) and then realized that
they actually were not in control.
How's that for admitting that SR can exist in living systems. It's just
pretty rare (and non-existent as a means of control -- I presume we know
what CONTROL means now?).
I'm not sending the above to Randy since he was not in on our SR discussion.
Richard S. Marken USMail: 10459 Holman Ave
The Aerospace Corporation Los Angeles, CA 90024
(310) 336-6214 (day)
(310) 474-0313 (evening)