The way the brain works (was Flame in terms of PCT)

Like Bjorn Simonsen, my attempts to incorporate PCT into my thinking
create many reorganizations in the way I consciously perceive the
world.

As I am presently seeing it, we each have a personal perception of
the world and our actions are based upon this perception (by
perception I'm meaning an internal model that is coded into the
neural network within our brains).

Difficulties arise because the world we are perceiving is subject to
continuous change, and to keep up with these changes we should be
continuously updating our internal model. This we cannot do very
accurately because some of the changes we are not aware of, or, we
are perceiving them incorrectly.

Differences become apparent to us when be carry out an action based
upon our perceived world (as it is represent by our internal model)
and the results are not those we are expecting.

To my mind, PCT is lacking here because it doesn't include the
concept of emotions. A mismatch (or a match) between expectations and
results are signalled to our consciousness through emotions. These
provide the prompts for us to update our internal model either by
obtaining more information or by reorganization of this internal
model - or - make attempts to change elements of the external world.

It is my contention that we actively pursue a strategy of updating or
reorganizing our internal model or changing elements in the external
world because emotions are prompting us to do so. This will continue
until these emotions tell us we can stop.

Another problem I have with PCT is in the way PCT visualizes how the
brain works. It seems to assume that it operates in much the same way
as control systems in system theory. This is with bundles of neurons
acting as functions, which pass messages to each other and regulation
is arranged through suitable feedback loops.

Work in neurology over the past decade has shown this not to be the
case. The brain functions in a far more efficient way, where the
bundles of neurons act like dynamic complex systems, each of which
can be triggered to go into different steady states.

This can be visualized by regarding bundles of neurons in the brain
as functions that provides many simultaneous results. Altering the
state of any of these bundles of neurons can change some of all of a
function's results (each result will be feeding into other
functions). Such changes can be brought about by external inputs.

Knowing the details of how neurons interact with each other, or about
the message paths around networks of neurons provides no way to
understand how they work because the changes of state (bifurcations)
are non linear.

As a non linearities cannot be dealt with logically (through a
control system within the brain), a signal has to be given - an
emotion - telling the conscious part of the brain that interaction
with the external environment is required.

If you are used to thinking of the brain and the central nervous
system in terms of control theory, this view of the way the brain
works will call for a massive reorganization of thinking at the
highest levels of organization. Most people will resist such a major
change.

Peter Small

Author of: Lingo Sorcery, Magical A-Life Avatars, The Entrepreneurial
Web, The Ultimate Game of Strategy and Web Presence
http://www.stigmergicsystems.com

···

--

From[Bill Williams 23 March 2004 5:40 AM CST]

I won't try to engage in a commentary upon all that Peter's post
contains, but I found the following to be of particular interest to
me.

First Peter says, in commenting upon his perception that PCT contains
inadequacies, that:

Differences become apparent to us when be carry out an action based
upon our perceived world (as it is represent by our internal model)
and the results are not those we are expecting.

But, what about the times that actions have the effect of bringing about
changes that _are_ however improbable not precisely what we were
expecting, but more or less what was expected, at least sufficiently like
what was expected that the perception _does_ match the reference level
after passing through the Reference Input Function. If successful control
does happen, might this be evidence that Peter's skepticism about the
theoretical description provided by control theory is exaggerated?

Second, Peter says,

As a non linearities cannot be dealt with logically (through a
control system within the brain), a signal has to be given - an
emotion - telling the conscious part of the brain that interaction
with the external environment is required.

As a former airline pilot, and also from time to time flight instructor,
I have observed non linearities being very successfully dealt with,
very large disturbances being dealt with by errors consistently
maintained at values near zero with nominal jitter, not only without
perceptible emotion, but entirely without the slightest hint of
conscious awareness if any effort to maintain control. This
experience would appear to be in direct contradiction to what
Peter is asserting.
p = (ke* ko *r + kd * d )/ (1 + ko* ke)

Bill Williams

···

From: "Peter Small" <peter@PETERSMALL.NET>

[From Bill Powers (2004.03.23.1050 MST);

For those with access to NASA TV, there's going to be a special news
conference at 2:00 PM Eastern time today about some sort of "scientific
discovery" by the Mars Rover Opportunity. The next news conference was
scheduled for Thursday, so it must be something interesting.

Peter Small (2004.03.23) --

As I am presently seeing it, we each have a personal perception of
the world and our actions are based upon this perception (by
perception I'm meaning an internal model that is coded into the
neural network within our brains).

One of the highest hurdles to cross in going from conventional views to the
PCT view is the reversal of causation that is required. I have seen two
references to my 1973 book that cited the title as "Perception: the control
of behavior".

In PCT behavior is not caused by or based on perception. To be precise, the
direction and amount by which behavior changes is caused by the difference
between the current state of perception and the "reference state," the
specification for the perception that is intended or desired. The net
result of this arrangement, which is called negative feedback control, is
that behavior changes in such a way as to bring the current state of
perceptions to a match with corresponding reference states. This happens
automatically even when there unpredictable disturbances. In short, this
kind of system controls its inputs, not its outputs.

Best,

Bill P.