SWITCH CONTROL - RKC

<Bob Clark (940527.1350 EDT)>

Mary Powers (940517) Subject: Misc, from Mary

In your post you referred to my post of 940516, in which I pointed out
that:

the control of these switches has four possible answers: a) the
reorganizing system; b) pre-existing higher level systems, such as
established programs; c) no control at all; d) an entity to be
studied and analyzed.

a) You agreed with my conclusion:

that changes in behavior resulting from operation of the
switch-controller do not involve reorganization.

No further comment is needed here.

b) I had suggested that:

Combinations of switches selected by an existing program can
provide the necessary control.

You commented:

this implies programs are stored and brought up as needed. A strong
implication of planned action. I don't think one ever uses the same
program twice, certainly not in detail, the environment never being
exactly the same. One may have a goal similar to previous goals,
and at a lower level some remembered general ways that succeeded in
the past, but these are subject to moment-by-moment corrections
according to present circumstance.

This is very close to my own concept of these operations. I
certainly agree with your first sentence -- and I think we agree that
people use "planned action." I fully agree with both of your next two
sentences. However the "moment-to-moment corrections according to
present circumstance" require at least some minimal control of the
switches. These observations do not answer either the question of
the nature of the switch-controller, nor of its relation to the
hierarchy.

Your comments do not address the following questions: what brings
programs up from storage "as needed;" how such a process is
accomplished; and how new programs can be created. Clearly these
actions do occur.

c) from above, "no control at all." In the discussion in my post, I
added:

Without some form of "switch controller," the switches remain
unchanging.

Your comment:

Seems to me it's the changing higher level reference signals that
"throw switches" - i.e. set goals - for lower level systems.
Including in and out of fantasy, imagination, memory - or dealing
with the "real" world.

This is essentially my own view, when there _is_ a switch controller.

This "c)" alternative considers the results when there is "_no
control at all_." I should have emphasized that: in the absence of
_any_ "controller," _none_ of the switches in the entire assembly can
be changed. Thus the higher level reference signals that "throw
switches" are also unchanging. As a result, the entire system
becomes essentially mechanical in the sense that its performance is
completely determined. It resembles mechanical devices such as
automobiles, telephones, computers, etc.

In the discussion in my post, I pointed out the changes in
terminology that result from assuming there is "no control at all."
Perhaps a better statement of these changes would be:

*To an observer, this would resemble an S-O-R system, with a complex
*form of "O."

*To a theorist, the "disturbance" of the Perceptual Control System
*would become the "stimulus" applied to the mechanical system.

*And the "output" of the control System would become the mechanical
*system's "response."

*Given the initial state of the system, its performance is completely
*predictable.

d) from above, "an entity to be studied and analyzed." The results of
your own "study and analysis" are in your comments:

What a wise creature the DME is. The basis on which it must select
switch positions implies an entire parallel multi-level control
system.

That would be taking the capabilities of the switch controller to a
maximum. Desirable, perhaps, if one is designing a maximum
performance system.

Instead, one can begin with seeking those minimal characteristics
that are necessary for the switch controller to do its switching. I
had noted:

the individual must be conscious

You ask:

and why conscious? Have you never experienced, say, reading a book
and realized you missed the last three paragraphs because you've
been thinking about something else? Not a conscious diversion at
all.

Yes, Mary, I have indeed had many such experiences. Distractions
involve a shift of attention with the switches changing to a new
combination. This new combination could continue, or be very brief
with a return to previous settings.

Consciousness is required only at the moment of changing the
switches. And conscious attention can shift from one area to another
with great rapidity. It is also true that one can be performing one
complex task while thinking about another. (Familiar example: Martin
Taylor's mental algebra while driving home.) Occasional monitoring is
necessary in case switches need to be changed.

Consider the opposite: unconsciousness. For example, sleep,
anesthetic, etc. In such a condition the switch controller is not
able to operate his switches.

To observe the operation of the switch controller may be difficult.
Memories of recent events must be examined.

Minimal requirements may be found in individual development.
Compared to an adult, a child or an infant, has very few memories to
be examined. And there are correspondingly few stored activities
that could be applied. [Note your interesting report of Ethan's
learning to walk on a rough surface.] But the switches are there, and
subject to the controller. The switching will often be unwise,
leading to additional experiences to be remembered.

Such exploration and experimentation are the basis of the
experimental method. They remain critically important throughout
life.

Regards, Bob Clark

From Tom Bourbon [940527.1350]

<Bob Clark (940527.1350 EDT)>

Mary Powers (940517) Subject: Misc, from Mary

In your post you referred to my post of 940516, in which I pointed out
that:

the control of these switches has four possible answers: a) the
reorganizing system; b) pre-existing higher level systems, such as
established programs; c) no control at all; d) an entity to be
studied and analyzed.

a) You agreed with my conclusion:

that changes in behavior resulting from operation of the
switch-controller do not involve reorganization.

No further comment is needed here.

b) I had suggested that:

Combinations of switches selected by an existing program can
provide the necessary control.

You commented:

Mary:

this implies programs are stored and brought up as needed. A strong
implication of planned action. I don't think one ever uses the same
program twice, certainly not in detail, the environment never being
exactly the same. One may have a goal similar to previous goals,
and at a lower level some remembered general ways that succeeded in
the past, but these are subject to moment-by-moment corrections
according to present circumstance.

Bob:

This is very close to my own concept of these operations. I
certainly agree with your first sentence -- and I think we agree that
people use "planned action." I fully agree with both of your next two
sentences. However the "moment-to-moment corrections according to
present circumstance" require at least some minimal control of the
switches. These observations do not answer either the question of
the nature of the switch-controller, nor of its relation to the
hierarchy.

For my "formal" paper in Wales, I'll be demonstrating a somple variation on
the theme of tracking -- what else. In the demo, at the start of each
quarter of the one-minute run, the time appears on the screen and a tone
sounds briefly. Nothing about those events "makes" anything else happen,
but a tracker can use them as time marks to demonstrate very clearly a
program level of control that selects reference perceptions to be controlled
by actions determined at lower levels. A person can easily repeat the same
program-selected sequence of reference perceptions and produce them, whether
or not the cursor is affected by a disturbance.

Mary, I believe people do this sort of thing very often and it does not
imply or require programmed actions; instead, the program is for a sequence
of _perceptions_. Bob, I agree that there is nothing in the demonstration I
just described that explains how and when the individual reference
perceptions first appeared, or how and why the program exists. But a simple
modification of the traditional single-loop PCT model that predicts
tracking so easily and accurately can duplicate the performance of a person
who selects a sequence of reference perceptions.

Your comments do not address the following questions: what brings
programs up from storage "as needed;" how such a process is
accomplished; and how new programs can be created. Clearly these
actions do occur.

Yes. And my simple demonstration does not address your questions, either.

We are about to use this sequential control task with some neurological
patients. The clinicians say it seems to have face validity, in that
suchpatients often have difficulty remembering and performing sequences
of "actions." (The clinicians are beginning to learn that the problem might
be one of perceptions, not actions.)

Later,

Tom