[From Bob Clark (930315.13:15 EST)]

Rick Marken (930307.1500)
You refer to my remarks about error, illustrated with a two level system.
My concern there was to point out the necessity for more than one level if
"error" is to be perceived.

My view of the "User" is stated more completely in my previous post, Bob
Clark (930301.1730). This is the view of the entity "riding around within
the organism." This is like the view of the driver of a car, or pilot of an
airplane. That is, he has little interest in the internal operations of
the machine -- as long as it does what he wants. He "controls" it by using
the levers, wheels, etc available. To him, it is pretty much a "S-R"
system. "Push" here -- "something moves" over there. If the "movement" is
not as desired, a different "push" is tried. When the User is familiar
with the machine's operation, he selects "Pushes" from his remembered
repertoire. Otherwise he uses a "trial and error," "experimental" method.
Of course if he has more complete information about the workings of the
machine (perhaps a remembered theoretical analysis), he can work out
solutions to unusual situations with less experimentation.

To me, this is an "Internal Viewpoint." This is what is available from the
"Inside." Note that this includes operation of the User's DME as it reviews
available decisions from the User's Memory.

Your discussion of "spilled wine" illustrates the difference between the
Viewpoints of the User and Observer -- as I define them. The User might
have intended the "spill" -- you'd have to ask him. The Observer would be
applying his own view of the "spilled wine event." I think this is
consistent with the view indicated in your remarks:

*But we know that very often a perception that is an error to one person
*is not an error at all to another. The spilled wine, for example, might
*be just what someone wanted to see -- that's why they knocked over the

The User is using his systems to bring his perceptions into agreement with
the conditions that he selected.

As you pointed out: all the Observer perceives is the spilled wine. (And
the actions of the Bystanders and the Spiller himself.) The Observer is
using his systems to observe and remember (at least briefly) what is
happening in the region to which he is paying attention.

*If the perception of spilled wine seems like an error to an observer, it
*must be because that perception deviates from some specification IN THE
*OBSERVER of what should be perceived. This seemingly obvious fact about
*perception is completely missed .......

Of course!! but people frequently get their "viewpoints" confused.

The Engineer is using his knowledge and skills to make the system he is
"building" (whether using hardware, software, or whatever) resemble the
"human" systems he is interest in. The Engineer doesn't care whether or
not the spilled wine is an "error." He observes that people make mistakes
and he is interested in building a multi-level control system that can
"perceive error" and correct it. An "Error" perceived by the system could
be a "Disturbance" originating "Outside" the System. An "Error" could also
originate inside the System in the form of a "Conflict." This form of
"Error" perceived by the system could result from incompatible requirements
derived from independent sets of "Goals."

Perhaps the following will clarify my earlier post, Bob Clark (930301.1730).

The Engineer's goal seems to be the construction, at least in principle, of
an assembly of hardware (or equivalent computer-cum-software) that performs
the same way that a human (or, perhaps, a simpler organism) does.

Some Engineers approach this in terms of levers, gears, pulleys, etc
arranged so that inputs ("disturbances"?) at certain locations result in
movements at other locations. By adding suitable "leading" terms (time
derivatives) and "lagging" terms (time integrals) these systems can be quite
effective for specified applications.

The PCT Engineer, if that is a suitable term, bases his design on the
properties of negative feedback control systems. These are combined into a
hierarchical structure, HPCT, assembled and modified to operate according
to his desires. The Engineer proceeds by selecting from his inventory of
memories, including physical and other principles, in order to bring his
proposed structure into correspondence with his view of human behavior.

The design might include "recording and playback" capability as well as
ability to "Reorganize" itself. In principle, these are both included in

To the Engineer an "Error" is a "Mistake" in his design. This "Error" is
revealed by an inconsistency between the performance of his "Engineered
System" and the performance of the system he is trying to imitate.


Bob Clark