[From Bill Powers (940812.0820 MDT)]
Bill Leach (940811.1644) --
The only question that this "answer" raises for me is that I doubt that
position is the only control loop that could be setting references for
things like tension and muscle length. As a quick "pops into mind"
type item, where does "shivering" come from? I see that this could be
rapid oscillation of position for one possibility.
As I experience shivering, it usually entails a rise in muscle tone
(i.e., the common-mode signals to opposing control systems are set to
large values). Perhaps the gains are set high enough to make the systems
unstable. When we shiver, our limbs are not limp!
But that's a detail. As you said in your excellent reply to Bob Clark,
such details may be beyond the requirements of the student (meaning us).
There's a temptation to start looking at secondary issues once the basic
organization is understood, and I think we may not be ready for that. A
really thorough model of the lower systems would require getting up to
speed on all the lore of physiology, of which there is a great deal.
Being lazy, I am hoping that a physiologist who is already acquainted
with that store of knowledge will take my primitive model of arm control
and turn it into a really grown-up model. In the meantime, I think we
should be content to say that we have a good understanding of the
general idea, and not try to get too fancy.
My concern is, I suppose, that I have not thought about the idea that
all muscular behaviour could be explained through a hiearchy that
always has position as a "superiour" level to tension type control.
There is a loop that I have played with a bit but not developed, the one
involving touch sensors, which also feed back directly to motor neurons.
In fact the arrangement in the spinal systems is extremely clever. If
the arm moves but the hand is in contact with some large object, the
position reference signals _automatically_ begin to play the role of
sensed-tactile-force control systems. With the touch sensors feeding
back directly to the motor neuron, the effect of a position error signal
(derived from muscle length) becomes the same as a touch reference
signal. There is no conflict, at that level, with the touch control
system, as far as I can see. We characterize the control systems in ways
convenient to the understanding, but the real thing is much more
sophisticated than our understanding of it.
One day, when the battle for survival of PCT is over, people will be
able to relax and really look into all these control loops without
worrying about having to justify what they're doing. There is yet a lot
left to learn about "spinal reflexes."
···
----------------------------------------------------------------------
Best,
Bill P.