Wallowing in reorganization

[From Bill Powers (930929.1445 MDT)]

Hal Pepinski (930929) --

I believe one side of us indulges in looking for gaps between
what you call our perceptions and our referents, all the way to
your ninth order, as opportunities to learn how to work better
with others.

Not referents -- reference signals, reference conditions,
reference perceptions. Things with reference to which we judge
our perceptions to be too little, just right, or too much. This
leaves "referent" free to be used in its original meaning: that
to which a term or symbol refers, as in "the referent of the word
'apple' is a round red thing you can eat." We already have too
few words for all the concepts to which we would like to point;
no sense in wasting the ones we've got.

In the PCT model there is a part of every control system at every
level that looks at the gap between perceptions and reference
signals: that is, between what we want to perceive and what we
are actually perceiving. This is the comparison function, or
comparator. The output of the comparator is the error signal that
indicates the size and direction of the gap. If the gap can be
closed by actions we already know how to perform, we simply
perform the action that will close it and all is well.

There may be situations in which the gap can't be closed. We
might, as you suggest, look at how we're working with others, and
evaluate that as being less good than how we could be working
with them. We have a reference picture of how we might work with
them, and a perception that differs from it. The difference could
simply be eliminated by changing our actions, but only if we had
already learned some lower-level behavior that would suffice to
eliminate it.

If we can't simply adjust our behavior to achieve the
improvement, the control system fails to work. The gap is not
closed. If this condition persists, something new has to be
learned: some new lower-level action, a new way of perceiving the
situation, or a new choice of reference signal (goal). Since
there isn't any already-existing control system that will provide
what is needed (if there were, it would have acted), we must
somehow modify or create a control system.

In PCT we call the process of modifying or creating control
systems "reorganization." This is a process separate from that of
the learned hierarchy of control which acts strictly as it has
been organized to act. The reorganizing system alters the
organization of the brain to create new systems. It is also a
control system, but its actions are not systematic. For want of a
better term, they are _random_. You could also term them
creative, experimental, or inventive.

Reorganization does not try to produce any particular
organization. The criteria with respect to which the reorganizing
system operates have nothing to do with the content of the
learned systems. To put this very simply, the reorganizing system
monitors the entire body and brain and perceives them as a set of
_intrinsic_ signals, signals that report the state of the whole
organism in critical regards. Each such intrinsic perception is
compared with an intrinsic reference signal, given by
inheritance, to yield an intrinsic error signal. Intrinsic error
signals indicate the deviation of the whole organism from a
particular state which evolution has defined as required for
survival. The sum of absolute values of all these individual
error signals is called "total intrinsic error". That is usually
what we mean when we say "intrinsic error." Total intrinsic error
determines the rate at which reorganizing events occur in the
parts of the brain that support the hierarchy of control systems.

The way the reorganizing system works is simple. When there is
intrinsic error, the output of the reorganizing system begins to
alter the basic organization of the learned hierarchy, altering
any and all components of the control systems. The rate of
reorganization increases and decreases as total intrinsic error
increases and decreases. The alterations themselves go in random

When a consequence of reorganization is such as to reduce
intrinsic error, reorganization slows; there are fewer
reorganizations per minute or hour or day. When intrinsic error
is eliminated (or drops below some threshold), the rate of
reorganization drops to its minimum, which may be zero or may be
some slow background rate. The result is that the last
organization that existed before intrinsic error was corrected
continues in effect.

This is the PCT alternative to reinforcement theory. It's what we
sometimes refer to as "E-coli-type" learning.

The reorganizing system is an concept meant to account for all
changes in the organization of the learned control systems. It
purports to model the mechanism underlying learning, creativity,
invention, and so on. As presented here, it is too simple a
system, and requires considerable elaboration to handle many
problems, such as the problem of directing reorganization to the
places where a problem exists instead of to perfectly good
control systems that don't need changing. We have several
proposals of this sort that make the reorganizing system
considerably more plausible. We have also conjectured, with some
reason, that the locus of reorganization follows awareness, and
that the process of making arbitrary changes is related to what
we experience as volitional or creative action.

While we speak of this mechanism as an "it", it is really "us."
The intent is not to pare down human experience to fit a
Procrustean theory, but to shape the theory to make a suitable
fit with human experience. The reorganizing system is meant to
bring into the model the phenomena we associate most closely with
creativity and invention, with doing things in new ways in an
experimental way that has no justification in the learned
hierarchy that already exists. These sorts of phenomena occur;
the reorganizing system is a way of trying to fit them into the
overall model, or fit the overall model to them.

So for instance factory management might take workers who
protest that there are more important ends to production than
increasing profits as an occasion to consider what other ends
production might serve, and whether they might supplant profit
increase in the manager's own mind. In the alternative model I
propose, one would at your tenth level of organization perceive
that control was being achieved by drawing out the disparity
between what the manager and his or her workers intend.

The tenth level (really 11th in the latest revision) of the
hierarchy would be concerned with maintaining a particular system
concept of the factory, in the mind of each participant. It would
be the reorganizing system that would respond to failures of
control by bringing up possible new ways of perceiving the
factory, new goals within old conceptions, new principles that
might yield the same old concept but by new means that cause less
error in all the participants. The process of reorganization is
started, automatically, by the misery that these failures of
control produce. Neither workers nor management would want to
change a system concept that resulted in all their goals being

Being random, reorganization does not guarantee improvement. But
if the result of a reorganization is more misery instead of less,
the rate of reorganization will go up; another reorganization
will follow immediately. If the result happens to be an
improvement, the next reorganization will be delayed; if the
improvement is sufficient, it may be delayed for a very long

"Drawing out the disparity," I presume, is where you would place
the deliberate creation of conflict (among others) in order to
produce a good end. But the problem is being caused by a conflict
of goals already; there is no point in, or no need to, exaggerate
it. According to our current thinking about reorganization, the
main thing that's needed in resolving conflicts is to bring both
sides of an already-existing conflict simultaneously into
awareness, which is another way of reading "drawing out the
disparity." This is an important aspect of individual
psychotherapy, where the primary problem is always some sort of
internal conflict, but with only one side of it in awareness.
Bringing any control process into awareness will tend to focus
reorganization on it, and it will begin changing. This is one
reason you don't want to become too acutely aware of the details
of any finely-honed skill. If it's not broke, don't start fixing

The only way to address issues like this is to wed spiritual,
non-material constructs to material patterns.

This is essentially what I am trying to do: to wed human
experience as it actually happens, without leaving out anything
important, to a material model of how the brain and nervous
system work. A lot of concepts that we think of as spiritual and
non-material are simply learned perceptions and function in a
perfectly mechanical way. But there are a few concepts such as
awareness and volition that seem inexplicable in any material
terms so far known. I would accept those as being "spiritual"
concepts, while insisting at the same time that they be part of
the model. I don't insist that PCT be confined to the concept of
a three-dimensional physical world as currently represented in
the physical sciences. I don't lightly violate physical
principles -- in fact I never do, on purpose. But neither do I
believe that the physical principles known in 1993 cover every
aspect of reality.

I think you explained yourself more clearly this time around. I
hope I have addressed some of the issues you raised, and some you
only implied.



Bill P.