[From Bill Powers (931110.1115 MST)]
Hans Blom (931110) --
I claim that you are a different person from me and that the
difference between us cannot be reduced to a different quality
of our receptors or actuators. I claim that the differences
between us must also depend on "tuning" of internal systems
that is based on our differing experiences, is somehow stored
in our memories -- conscious or not -- and is available in the
processing that goes on between perceptions and motor output.
I agree that we are different and that many of our differences
arise from different experiences with the world. I agree that
higher levels of organization exist in the brain, and that they
go beyond "qualities" of receptors and effectors.
If you are saying that behavior is produced by perceptions acting
on our inner organization to cause motor outputs, however, then I
totally disagree: PCT is about quite a different relationship
between perceptions and outputs.
Please define kinesthetic control of walking in the dark in
terms of feedback processes only.
Walking is a process of varying reference signals for control
systems that make perceptions of joint angle velocity follow a
perceived oscillating temporal pattern. Velocity control is
accomplished by these control systems' sending varying reference
signals to systems concerned with controlling signals
representing static joint angle; those work by sending varying
reference signals to systems controlling muscle length signals
(from muscle spindles) and tendon stretch signals (from Golgi
receptors).
At the same time, other control systems at a comparable level are
adjusting the sensed leg positions over several strides to
maintain an average perception of foot pressure within a small
region under the center of gravity (how that is computed I don't
know), and also to control vestibular signals indicating inertial
measures of body-head angular accelerations. The direction of
walking is adjusted by varying reference signals for sensed twist
at the ankles and hips.
All of these kinesthetic control systems work the same in the
light or in the dark. Systems concerned with controlling visual
perceptions operate by sending reference signals to the top level
of these kinesthetic control systems, specifying the desired
states of the perceptions of walking speed and turning rate, but
not specifying the details.
This is my current understand of how these systems work. At the
lowest levels (spinal and brainstem reflexes) the control-system
organization can be matched to neural circuitry very closely, in
detail. At higher levels evidence is less direct.
We don't just stride on ahead when the lights go out. We go
into a drastically different mode of behavior.
Your remarks are demonstrably false. Everyone can do the
following easy experiment. Take a walk on a dark night in a
predictable environment where the road is even (no kinesthetic
feedback) and there is no noise (no auditory feedback).
Well, sure they're false if you change the example entirely. I
was following your example of walking in a bedroom full of
furniture with the lights out. You have changed the conditions so
that no caution is required, no path around nearby obstacles must
be negotiated, and no specific final destination is sought. Other
than that, I suppose you could claim that the situation is
exactly the same.
There is nothing to keep you from walking through the dark
bedroom in exactly the same free and confident way you walk when
you are not trying to get to a particular place and know that
there are no obstacles in the way. Nothing, that is, but your
intention of avoiding running into things.
I was considerably startled by your statement that walking on a
level road implies that there is no kinesthetic feedback. What on
earth do you think "feedback" is? Whenever you walk, perceptual
signals arise from every muscle, every tendon, and every hip and
leg joint, not to mention the pressure sensations from the soles
of your feet (not strictly kinesthetic, but certainly somatic).
Hundreds of thousands of perceptual signals are continuously
varying while you walk; without the feedback from those signals
you wouldn't be able to stand up, much less walk. You would be
like those deafferented animals, lurching and staggering and
crawling about with a pitiful lack of motor skills.
"Feedback" is not just "sensory input." Sounds are feedback only
if you make them yourself. Feedback is an effect of a signal or
variable ON ITSELF via a closed loop. I doubt very much that you
could walk on a road at night without generating any auditory
feedback -- the road would have to be surfaced with a nice thick
carpet for that to happen.
A typical lonely country road without a lining of trees (which
would reflect the sounds of your steps) will do.
Are you saying that if there are no trees to reflect sound you
can't hear your own footsteps? I can hear myself walking even on
a carpet. Are you somewhat deaf? (Actually, I'm a little deaf, so
you must be in even worse shape).
I live on a lonely country road, and I can tell you that when I
haul the garbage out after dark, I have considerable problems
with walking. Where I live, "smooth" and "country road" don't go
together.
Walk at a brisk, steady pace. Then let your eyes go into a
"reverse blink" pattern; that is, close them for about one
second, then open them for a very, very brief period of time. I
predict that you will feel no need to slow down.
If I'm allowed sampled visual feedback, that's almost as good as
continuous feedback, especially if I can see that there's nothing
in my way and I'm not trying to maintain a steady course across a
rapidly-varying wind, or walking along on a narrow plank. The
kinesthetic control systems work the same way as ever; all that's
different is that the visual systems can make an error correction
only at intervals (altering the reference signals specifying rate
and direction of walking for the kinesthetic control systems).
See how far you can extend the periods between "reverse
blinks". You will be surprised. No slowing down, no groping, no
searching for glimmers of light.
Not surprised a bit. Everyone has tried this and knows how it
works. You don't need to open your eyes at all if the nearest
obstacle is 100 meters away and you know the surface is flat.
You're just walking under kinesthetic control. Try walking in a
straight line with your eyes shut and see how long it remains
straight.
The mode of control changes not a bit.
You know better than that, Hans. A sampled control system HAS to
operate differently if the sample interval changes from 1 second
to 5 seconds. If the system is stable for a 5-sec interval, it
will be more sluggish than necessary with a 1-sec interval. If it
is operating with optimum filtering for a 1-sec interval, it will
become unstable with a 5-sec interval. All you're saying is that
you didn't notice any difference (assuming you actually did this,
and are not just concocting an imaginary example to support your
theory).
You can't tell much about the mode of control unless there are
disturbances acting. Only when you're shoved off course, or when
your destination is moved, can you see how rapidly you recover
the correct course, and that's what tells you if something about
the control system has changed.
Without applying disturbances, or observing the effects of
natural ones, you can't tell whether ANY behavior is open-loop or
closed-loop. Either model is just as plausible until you apply
the test. A control system will alter its actions to oppose the
effects of a disturbance; an open-loop system won't even know
that a disturbance happened.
Maybe I should emphasize that point: you can't tell from watching
any system in a disturbance-free environment whether it's a
control system or an open-loop system. It could be either one;
both kinds of system would behave in the same way. You can only
tell the difference when a disturbance alters the outcome. The
open-loop system is the one that does nothing about the
alteration.
This clarifies something in my mind if not yours. You and Martin
Taylor have both been insisting that in some circumstances,
environments are predictable enough that open-loop systems could
operate properly. But what is your evidence that the systems you
observe are actually operating open-loop? The only evidence that
would matter would come from seeing what happens when the
environment is disturbed. And if you're confining your definition
of open-loop behavior to behavior in undisturbed environments,
you have no way of knowing that you're not looking at a control
system. To prove that you're talking about an open-loop system,
you would have to show that a disturbance would NOT be resisted.
And if you DO see that disturbances are resisted, it's really not
acceptable to say, "Oh, it WAS an open-loop system before we
tested it, but as soon as the disturbance was applied it turned
into a control system." Sometimes it seems to me that some of you
are getting close to that kind of last-ditch rationalization. You
know, everything falls upward except when someone looks at it,
and then it falls downward.
ยทยทยท
--------------------------------------------------------------
Best,
Bill P.