Attractors & Evolution

[Martin Taylor 2004.05.12.01.05]

[From Peter Small (2004.05.11)]

My impression is that Martin hasn't yet cottoned on to the the
concept of attractors as they are applied to biological systems. His
description of attractors is limited to describing patterns of
attractors as they would appear on a computer screen as a result of a
simple mathematical formula. There is no attempt to relate this to
how they might be represented in the brain and how the concept can be
applied to memory and perceptions.

From what I've read of Martin's descriptions, he appears to confuse
attractors with catastrophe theory - which is a totally different
concept.

It's a good think you wrote this on CSGnet, not on the ECACS Forum,
because its innacuracy and tendency toward personal insult would
probably have got it removed to the Roughneck Room, had it been
presented over there.

I don't really think you can have read much of what I have written.
Or if you did, you put a very peculiar interpretation on it.

Chaos theory is about the iteration of a function that repeats and
has a variable, which is the result of its previous iteration. i.e.

Phase 1 = f(x) = a
Phase 2 = f(a) = b
Phase 3 = f(b) = c
Phase 4 = f(c) = d
etc., etc., etc.,

Useful though it is for didactic purposes, that really is a terribly
limited view of chaos, for two reasons. One is that it treats of
discrete time, and two is that there are many different kinds of
chaos, depending on how many of the Lyapunov exponents have what
sign. The only thing that is really relevant here is that chaos does
require a feedback of effect such that a change in teh value of a
variable in some way affects its future value.

By changing the initial values of the function, it will have a
different trajectory and may have a different attractor in this space.

Badly stated, but not technically incorrect. The different attractor
always existed in the dynamical landscape of the function. Different
initial values may have resulted in an orbit in a different attractor
basin, but that basin always coexisted with the one in which the
original initial values placed the function.

A dynamical system can have multiple trajectories and multiple
attractors. The conceptual breakthrough is to see how these
attractors can be regarded as addressable memories in the phase space
- similar to the way the bit space of a computer memory can be
addressed.

This much is true. But, as Bill P suggested a few days ago, it does
imply categorical perception. Even though Bill didn't understand the
dynamical background very well, he was correct in that much.

Here's a chunk from "Thoughts on the Edge of Chaos" that attempted to
introduce a similar idea, but in a more precise way:

···

-----------------------
  Replication is the central concept of our view of cognition. It is
through replication that tenuous concepts attain stability, that
concepts evolve to match environmental influences, that learning
happens and that concepts fuse to create novel insights. It is also
through replication that the chaotic regime necessary for attentive
perception can be developed.

It is not possible for an isolated system to replicate. Hence
replication is a property that occurs only in part of an isolated
system; moreover it is in a part through which there is an energy
flow, so that the sub-system (or module) can acquire negative entropy
from the input and pass positive entropy to the output. A replicating
module maintains itself at a particular energy level, averaged over
an entire replication cycle.

The replication of a module does not mean that after replication
there exist two such modules; the act of replication may involve the
destruction of the original module. What it means is that the
existence of the module in its particular environment will result in
the later emergence of a more or less faithful copy of the original
module. The original module may still exist, in which case the
environment for both is different from that which permitted the
replication. The new environment may or may not permit the further
replication of the two modules. Replication is a function not only of
the module, but of the environment in which the module is found. The
congeniality of the environment is critical.

Replication does not always mean duplication, though if the process
is to be stable, the replicated version should be very like the
original. The term "replication" is used in the sense that x n = f (
xn-1) is a replicating function. f need not be the identity function,
but if it is to be useful in a somewhat stable system, successive
applications of f should not move x n too far from x n-p after an
initial transient. In other words, the attractor of a useful
replicating function should not be at infinity.

If the attractor of f for a particular x0 is a point, then
replication eventually becomes duplication. If the attractor is a
cycle, xn is identical to xn-p for some value of p. This condition is
like the generational cycle of living organisms. The baby is unlike
the child it becomes, which is unlike the adolescent, which is unlike
the adult, which is unlike the baby born of the adult, although the
new baby is quite like the original one that became the adult. (Of
course, the new baby's genetic structure is different, but that is
beside the point for this example.) The replications concerned with
the functions of life generally approximate cyclic attractors rather
than points. We shall argue, however, that the attractors may well be
strange, looking sometimes as if they were cyclic, and at other times
not. Such nearly periodic behaviour is typical of strange attractors.
Long periods of apparently cyclic behaviour are separated by short
periods of rapidly changing behaviour. The non-quasi-cyclic intervals
perhaps correspond to the moments of species shift in punctate
evolution, to moments of insight in cognition, or to paradigm shifts
in the social structure of science.

Replication implies that so long as the environment is conducive, the
same (or similar) patterns will recur. The provision of an
environment favourable to replication is essential to the stability
of forms, since it is through replication that a damaged form may be
repaired: somewhere there exists a template derived from the original
form or from some part of the cycle that produced the damaged form,
and from this template the damaged parts can be replaced by correct
parts. In DNA replication, the template for each half of the double
helix is the other half, and because the halves are complements of
each other rather than duplicates, they are unlikely to be damaged
similarly by external events. It is probably a fairly general
statement that an effective replicating environment does not involve
the use of copies of the thing being replicated; rather it may use
complementary copies, algorithms that produce elements of the thing,
micro-structures that favour the recreation of the original in its
absence, and so forth. Such environmental structures are unlikely to
be subject to exactly the same damaging influences as the thing being
replicated, and thus form a more stable memorial milieu than simply
the provision of multiple copies of everything.

-------------------------------
Referring to Weiner:

Isn't this 1940's way at looking at the brain somewhat similar to the
underlying concept of PCT?

Weiner's work is ordinarily regarded as one of the historical
antecedants of PCT, yes. Are you implying that because it is 1940's,
it is therefore wrong, and therefore so is PCT?

Martin

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.01.05]

[From Peter Small (2004.05.11)]

My impression is that Martin hasn't yet cottoned on to the the
concept of attractors as they are applied to biological systems. His
description of attractors is limited to describing patterns of
attractors as they would appear on a computer screen as a result of a
simple mathematical formula. There is no attempt to relate this to
how they might be represented in the brain and how the concept can be
applied to memory and perceptions.

From what I've read of Martin's descriptions, he appears to confuse
attractors with catastrophe theory - which is a totally different
concept.

It's a good think you wrote this on CSGnet, not on the ECACS Forum,
because its innacuracy and tendency toward personal insult would
probably have got it removed to the Roughneck Room, had it been
presented over there.

Well, well, well, Martin. You can criticize the views of others and
tell them they are wrong, but, when somebody does this to you, you
bristle up and accuse them of personal attacks. Are you omniscient?

Martin wrote:

Weiner's work is ordinarily regarded as one of the historical
antecedants of PCT, yes. Are you implying that because it is 1940's,
it is therefore wrong, and therefore so is PCT?

Martin

Non sequiturs are not honest forms of discussion. I'm not implying
that PCT is wrong because it is based upon ideas born in the 1940's,
I'm saying that PCT is wrong because it is based upon out-dated
concepts that have not been substantiated by the last thirty years of
neurological research. Better concepts have superceded the concepts
upon which PCT is based. Weiner's paper explains this quite well, so
I presume you haven't bothered to read it.

The PCT paradigm is based upon the premise that there is some kind of
error signal generated between an internal reference perception and a
sensed perception of the outside world. This is totally Mickey Mouse
stuff.

The current view of most neurologists is that the concept of
attractors explains how the brain can reduce the degrees of freedom
in the vast amount of information arriving at our senses to a
manageable amount. In other words, attractors enable the brain to
divide the information up into discrete chunks of meaning.

These chunks of meaning are the inputs that activate a myriad of
attractors to construct what we term a perception. PCT would have it
that these perceptions are compared to some form of reference
perception. In terms of information theory, the PCT view implies a
measurement of the manning distance between the observed and the
reference.

The problem with this view is that there is no evidence of this
happening and in theoretical terms it is impossible for there to be a
reference perception for observed perceptions to be compared with.

The current view is that discrete elements of meaning are extracted
from the signals arriving at the senses and then used to construct a
perception that includes links to various mechanisms in the brain
that trigger sensations such as awareness and emotion. This building
process is a highly selective and modulated process and is constantly
changing. There are no comparisons made with past perceptions. In
fact there isn't any concept of comparison included at all.

I cannot hope to explain this conceptualization fully in a short post
to a forum, but I have provided several references to papers that
provide starters for these modern concepts. You can put your head in
the sand and ignore them if you wish, but this will further isolate
PCT from mainstream thinking - putting it further into disrepute.

If PCT is a realistic explanation of cognition, wouldn't it have been
adopted for use in AI systems? It hasn't, but the concept of using
attractors for cognition in AI is steadily gaining ground. May I draw
your attention to a paper by Dr. Peter Henningsen entitled
"Complicity and the Brain: Dynamics in Attractor Space"

This paper will perhaps provide a few clues as to how PCT is totally
incompatible with the present day conceptualization of how the brain
works.

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

···

--

[Martin Taylor 2004.05.12.11.12]

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.01.05]

[From Peter Small (2004.05.11)]

My impression is that Martin hasn't yet cottoned on to the the
concept of attractors as they are applied to biological systems. His
description of attractors is limited to describing patterns of
attractors as they would appear on a computer screen as a result of a
simple mathematical formula. There is no attempt to relate this to
how they might be represented in the brain and how the concept can be
applied to memory and perceptions.

From what I've read of Martin's descriptions, he appears to confuse
attractors with catastrophe theory - which is a totally different
concept.

It's a good think you wrote this on CSGnet, not on the ECACS Forum,
because its innacuracy and tendency toward personal insult would
probably have got it removed to the Roughneck Room, had it been
presented over there.

Well, well, well, Martin. You can criticize the views of others and
tell them they are wrong, but, when somebody does this to you, you
bristle up and accuse them of personal attacks. Are you omniscient?

I didn't complain about being told I'm wrong. That happens a lot, and
somethimes the critic is right, sometimes not. What I complained
about was that your criticism was based on fantasy, rather than on my
writings or the positions I have claimed.

Criticism is the way science advances, but it works a lot better if
the position being criticised has in fact been advanced. To build
your own straw man and give it someone else's name is personal
attack, not scientific criticism.

Martin

[Martin Taylor 2004.05.12.11.15]

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.01.05]

Weiner's work is ordinarily regarded as one of the historical
antecedants of PCT, yes. Are you implying that because it is 1940's,
it is therefore wrong, and therefore so is PCT?

Martin

I'm saying that PCT is wrong because it is based upon out-dated
concepts that have not been substantiated by the last thirty years of
neurological research. Better concepts have superceded the concepts
upon which PCT is based. Weiner's paper explains this quite well, so
I presume you haven't bothered to read it.

As it happens, I got into the whole business of control theory
because of reading both "Cybernetics" and "The Human Use of Human
Beings" as an undergraduate in the early 50s. I don't know what paper
of Weiner's you refer to, but those were part of my intellectual
foundation, and in part why PCT appealed to me when I finally
understood that it was truly a functional description rather than a
descriptive metaphor. They were why I nearly went into control theory
as my graduate work.

I'm glad I took psychology instead, but that background has always
coloured my approach to psychological issues. As a psychologist, my
research has mainly been in perception and psycholinguistics, which
involves the construction of cognitive concepts. I don't know when or
why you first got involved in the use of attractor systems for the
development of cognitive constructs. For me, the notion became more
or less fundamental perhaps 15 years ago, so that it is now more a
natural tool in my toolbox than a great epiphany that I have to
propagate to the unwahsed masses.

Whether the specific elaborations of PCT are correct or need
modification is anothe matter. The fundamental absolute necessity of
PCT as the basis for all life follows from basic physics. If that
changes, then maybe you might have an argument as to whether PCT is
wrong.

The PCT paradigm is based upon the premise that there is some kind of
error signal generated between an internal reference perception and a
sensed perception of the outside world. This is totally Mickey Mouse
stuff.

Good point. Not easily refuted. Especially since Mike Eisner is very
touchy about who uses Mickey's image.

The current view of most neurologists is that the concept of
attractors explains how the brain can reduce the degrees of freedom
in the vast amount of information arriving at our senses to a
manageable amount. In other words, attractors enable the brain to
divide the information up into discrete chunks of meaning.

Agreed. We wrote something like that 13 years ago in the paper for
BBS that was superseded by Freeman's.

These chunks of meaning are the inputs that activate a myriad of
attractors to construct what we term a perception. PCT would have it
that these perceptions are compared to some form of reference
perception.

Yes. Although you have to be a bit careful when you talke about "a
myriad of attractors", because you are talking about a view based on
modularization of the state space of the brain. The brain state is in
only one attractor at a time, but there are amyriad of
quasi-independent views in which the substates seem to be in their
own attractor basins. They aren't, really, so you have to be careful
that the use of English in this way doesn't mislead your thinking.

In terms of information theory, the PCT view implies a
measurement of the manning distance between the observed and the
reference.

No. I assume "manning" is a typo, but whether it is or not, no
measure of distance is required. It can be useful, but it's not
necessary.

The problem with this view is that there is no evidence of this
happening and in theoretical terms it is impossible for there to be a
reference perception for observed perceptions to be compared with.

That needs explanation. As a flat statement, it's hard to evaluate.

The current view is that discrete elements of meaning are extracted
from the signals arriving at the senses and then used to construct a
perception that includes links to various mechanisms in the brain
that trigger sensations such as awareness and emotion.

So far, so good.

This building
process is a highly selective and modulated process and is constantly
changing. There are no comparisons made with past perceptions. In
fact there isn't any concept of comparison included at all.

That may be a significant omission, but it doesn't mean that
comparisons aren't made. There are lots of ways in which they can be
done, not least of which is the location of a state point within a
_pre-existing_ system dynamic.

Yes, there's an evolutionary (stigmergic) relation between the system
dynamic and the system behaviour, but at any one moment, the state is
behaving in a dynamic that is already there.

I cannot hope to explain this conceptualization fully in a short post
to a forum, but I have provided several references to papers that
provide starters for these modern concepts.

For which we are grateful.

If PCT is a realistic explanation of cognition, wouldn't it have been
adopted for use in AI systems?

Idon't think anyone has argued that PCT is "an explanation of
cognition". If anyone provides an explanation of cognition, PCT ought
to be able to incorporate it, no matter what the mechanism. PCT is
about CONTROL of perception, not about how perception is generated.

It hasn't, but the concept of using
attractors for cognition in AI is steadily gaining ground. May I draw
your attention to a paper by Dr. Peter Henningsen entitled
"Complicity and the Brain: Dynamics in Attractor Space"

Complicity and the Brain: Dynamics in Attractor Space

This paper will perhaps provide a few clues as to how PCT is totally
incompatible with the present day conceptualization of how the brain
works.

I'll check it out, and see if I agree with you.

Martin

[From Bruce Nevin (2004.05.12 11:40 EDT)]

I am going to be very presumptuous and speak for a field of which I have
always been a marginally contributing member.

PCT is not about physiology. It is about control of perceptual input. The
central claim is that behavior is only and always the means to control
sensory input. B:CP talks about physiology to indicate that the model of
control is plausible, but of course it only refers to literature published
before 1973.

Peter, Marc, and others with expertise in biology, neurophysiology,
neurochemistry, you chide us for our limited and outdated knowledge of
these fields. If you can demonstrate some other model in which the
appearance of negative feedback control of perceptual input is a mere
appearance, an illusion, a side effect, more power to you. If you can
demonstrate that behavior is something other than means of controlling
perceptual input, we would be a most attentive audience.

But I don't think that is your aim. I think that what you want to do is to
demonstrate other neurological and biochemical mechanisms that accomplish
control of input, in lieu of outdated views assembled by us amateurs. If
you can do this, more power to you. You will have our rapt attention.

But it's not enough merely to contrast these mechanisms to our incorrect
and outdated ideas about neurology, etc. We will gladly and easily give
those up when you show how the actual neurological and biological
mechanisms accomplish control of sensory input. To do this, you have to
identify some control phenomena, that is, you have to pick an instance of
control and identify what perceptual variable is being controlled, and you
have to have data about the state of that variable and the effects of
everything that influences that state -- disturbances as well as behavioral
outputs. Then you construct a generative model that replicates those data
with extreme fidelity. Your knowledge of biological mechanisms can inform
the construction of the model, or it can demonstrate the plausibility of
the control model.

The phenomenon of control raises more questions than we can possibly answer
in a lifetime, even if we all work on it full time. It defines a field for
scientific work. We're doing the best we know how. Help us to do better.

        /Bruce Nevin

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.11.12]

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.01.05]

[From Peter Small (2004.05.11)]

My impression is that Martin hasn't yet cottoned on to the the
concept of attractors as they are applied to biological systems. His
description of attractors is limited to describing patterns of
attractors as they would appear on a computer screen as a result of a
simple mathematical formula. There is no attempt to relate this to
how they might be represented in the brain and how the concept can be
applied to memory and perceptions.

From what I've read of Martin's descriptions, he appears to confuse
attractors with catastrophe theory - which is a totally different
concept.

It's a good think you wrote this on CSGnet, not on the ECACS Forum,
because its innacuracy and tendency toward personal insult would
probably have got it removed to the Roughneck Room, had it been
presented over there.

Well, well, well, Martin. You can criticize the views of others and
tell them they are wrong, but, when somebody does this to you, you
bristle up and accuse them of personal attacks. Are you omniscient?

I didn't complain about being told I'm wrong. That happens a lot, and
somethimes the critic is right, sometimes not. What I complained
about was that your criticism was based on fantasy, rather than on my
writings or the positions I have claimed.

Criticism is the way science advances, but it works a lot better if
the position being criticised has in fact been advanced. To build
your own straw man and give it someone else's name is personal
attack, not scientific criticism.

Martin

Then perhaps you'd tell me where you have written about where and how
attractors are used in the brain and how they relate to the way in
which perceptions are manifested.

Your position as I understand it is to envisage some kind of error
between a real and an ideal perception. I'm saying that I do not
believe this position is tenable in light of what is now known about
neurodynamics.

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 Rick Marken (2004.05.12.0915)]

Bruce Nevin (2004.05.12 11:40 EDT)--

I am going to be very presumptuous and speak for a field of which I have
always been a marginally contributing member.

I think you should be this presumptuous more often, Bruce! This is the best
piece I've ever seen you post. Of course, that's because you say _exactly_
what I would have said. Which doesn't necessarily mean you are right. But it
almost certainly does;-)

Very nice post, Bruce.

Best

Rick

···

---

PCT is not about physiology. It is about control of perceptual input. The
central claim is that behavior is only and always the means to control
sensory input. B:CP talks about physiology to indicate that the model of
control is plausible, but of course it only refers to literature published
before 1973.

Peter, Marc, and others with expertise in biology, neurophysiology,
neurochemistry, you chide us for our limited and outdated knowledge of
these fields. If you can demonstrate some other model in which the
appearance of negative feedback control of perceptual input is a mere
appearance, an illusion, a side effect, more power to you. If you can
demonstrate that behavior is something other than means of controlling
perceptual input, we would be a most attentive audience.

But I don't think that is your aim. I think that what you want to do is to
demonstrate other neurological and biochemical mechanisms that accomplish
control of input, in lieu of outdated views assembled by us amateurs. If
you can do this, more power to you. You will have our rapt attention.

But it's not enough merely to contrast these mechanisms to our incorrect
and outdated ideas about neurology, etc. We will gladly and easily give
those up when you show how the actual neurological and biological
mechanisms accomplish control of sensory input. To do this, you have to
identify some control phenomena, that is, you have to pick an instance of
control and identify what perceptual variable is being controlled, and you
have to have data about the state of that variable and the effects of
everything that influences that state -- disturbances as well as behavioral
outputs. Then you construct a generative model that replicates those data
with extreme fidelity. Your knowledge of biological mechanisms can inform
the construction of the model, or it can demonstrate the plausibility of
the control model.

The phenomenon of control raises more questions than we can possibly answer
in a lifetime, even if we all work on it full time. It defines a field for
scientific work. We're doing the best we know how. Help us to do better.

      /Bruce Nevin

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

From[Bill Willliams 12 May 2004 12:10 PM CST]

[From Peter Small (2004.05.12)]

Well, well, well, Martin. You can criticize the views of >others and tell

them they are wrong, but, when >somebody does this to you, you bristle up
and accuse >them of personal attacks.

Some time ago, I lost all interest in whatever it is that you are attempting
to say. At that time I said that I would be content to leave it to Martin
to figure out if there was anything to what you are saying. I didn't myself
think that it was worth Martin's time, but Martin has this very persistent
irenic streak, and if he chooses as a matter of policy to approach
discussion in the way that he does, then that is his choice and perhaps a
good choice. But, you have managed to exhaust even Martins very
considerable fund of good will.

And, even Martin after alll does have a fuse.

Bill Williams

Phil Runkel replying to Bruce Nevin's message of 2004.05.12 11:40 EDT:

Bruce, thanks very much for your clear and concise contribution. I sign
on. --Phil

Martin Taylor 2004.05.12.1657]

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.11.12]

[From Peter Small (2004.05.12)]

[Martin Taylor 2004.05.12.01.05]

[From Peter Small (2004.05.11)]

My impression is that Martin hasn't yet cottoned on to the the
concept of attractors as they are applied to biological systems. His
description of attractors is limited to describing patterns of
attractors as they would appear on a computer screen as a result of a
simple mathematical formula. There is no attempt to relate this to
how they might be represented in the brain and how the concept can be
applied to memory and perceptions.

From what I've read of Martin's descriptions, he appears to confuse
attractors with catastrophe theory - which is a totally different
concept.

It's a good think you wrote this on CSGnet, not on the ECACS Forum,
because its innacuracy and tendency toward personal insult would
probably have got it removed to the Roughneck Room, had it been
presented over there.

Well, well, well, Martin. You can criticize the views of others and
tell them they are wrong, but, when somebody does this to you, you
bristle up and accuse them of personal attacks. Are you omniscient?

I didn't complain about being told I'm wrong. That happens a lot, and
somethimes the critic is right, sometimes not. What I complained
about was that your criticism was based on fantasy, rather than on my
writings or the positions I have claimed.

Criticism is the way science advances, but it works a lot better if
the position being criticised has in fact been advanced. To build
your own straw man and give it someone else's name is personal
attack, not scientific criticism.

Martin

Then perhaps you'd tell me where you have written about where and how
attractors are used in the brain and how they relate to the way in
which perceptions are manifested.

I had hoped you had read "Thoughts on the Edge of Chaos" early in our
electronic acquaintance. Here is the URL again:
<http://www.mmtaylor.net/Academic/Thoughts.chaos/Thoughts.chaos.1.html&gt;\.
It isn't much, as the paper was aborted when Freeman's very similar
one appeared in BBS, with real olfactory data, which we would never
have had. I think our paper was wider ranging than his, but there was
no way it would have been publishable. Nevertheless, you can see the
conceptual linkages in the chart on page 1. having seen the chart, I
defy you to say again:

"My impression is that Martin hasn't yet cottoned on to the the
concept of attractors as they are applied to biological systems. His
description of attractors is limited to describing patterns of
attractors as they would appear on a computer screen as a result of a
simple mathematical formula. There is no attempt to relate this to
how they might be represented in the brain and how the concept can be
applied to memory and perceptions."

But that wasn't the core of my complaint. It was your assumption that
I might be so naive as to confuse chaos with catastrophe. That was
quite an insult, and it seemed to be deliberate. I didn't appreciate
it.

Martin

[From Peter Small (2004.05.13)]

[Martin Taylor 2004.05.12.1657]

I had hoped you had read "Thoughts on the Edge of Chaos" early in our
electronic acquaintance. Here is the URL again:
<http://www.mmtaylor.net/Academic/Thoughts.chaos/Thoughts.chaos.1.html&gt;\.

<snip>

But that wasn't the core of my complaint. It was your assumption that
I might be so naive as to confuse chaos with catastrophe. That was
quite an insult, and it seemed to be deliberate. I didn't appreciate
it.

Martin

Martin,

It was from this paper that I formed the impression you were
confusing catastrophe theory with chaos

For example, on page
http://www.mmtaylor.net/Academic/Thoughts.chaos/Thoughts.chaos.3.html

You write:

"Cusp catastrophes will be shown to be important in learning, and to
provide a route between the chaos of alertness on the one hand, and
the stability of categoric perception or the precision of logical
thought on the other."

You then go on to infer that the switch between attractor basins is
explained by the fold of a catastrophe function.

From page
http://www.mmtaylor.net/Academic/Thoughts.chaos/Thoughts.chaos.6.html

you write:

···

---
"How can a semi-chaotic system perform logical operations with a
reasonable chance that the result will be predictable? The key is to
transform the state space in such a way that the boundary regions of
uncertainty become negligible in size as compared to the stable
attractor basins of the various categories. In place of the
single-valued state space, replace the boundary regions with a fold
catastrophe surface, as shown in Fig 2. When the underlying (covert)
state is in the region covered by the fold, the overt or externally
visible state is on one of the branches of the fold. Which branch it
is on depends on its history. If the underlying state drifts into the
region of the fold, the overt state stays on the contiguous branch,
and does not move to the other branch until the underlying state
drifts off the other side of the fold region. At that point, the over
state shifts abruptly to the other branch, and will not return to the
original branch until the underlying state drifts completely across
the fold region.

The bifurcation apparent in the transition across the cusp from a
single-valued function to a fold catastrophe will be a recurrent
theme in the different views of cognition that are coordinated in
this paper. It is a general property of the approach to chaos induced
by some kind of stress..."
---

In this paper, which is an explanation of chaos, there is a long
description of catastrophe theory. Yet it plays no part at all in
chaos and attractor theory.

Okay, so the paper was written several years back, at a time when
catastrophe theory was in vogue, but several times in recent posts
relating to attractors and chaos you have referred to catastrophe
events such as the instability of sand hills.

A consequence of this is your constant insistence that only one
attractor can exist in the brain at any one time. This is not true
because attractors in the brain are made up as a combination of many
attractors interacting with each other.

Anyway, the overall impression I get is that you are somehow mixing
the concept of catastrophe theory with the concept of chaos theory.

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 Peter Small (2004.05.13)]

[From Bruce Nevin (2004.05.12 11:40 EDT)]

I am going to be very presumptuous and speak for a field of which I have
always been a marginally contributing member.

PCT is not about physiology. It is about control of perceptual input. The
central claim is that behavior is only and always the means to control
sensory input. B:CP talks about physiology to indicate that the model of
control is plausible, but of course it only refers to literature published
before 1973.

Peter, Marc, and others with expertise in biology, neurophysiology,
neurochemistry, you chide us for our limited and outdated knowledge of
these fields. If you can demonstrate some other model in which the
appearance of negative feedback control of perceptual input is a mere
appearance, an illusion, a side effect, more power to you. If you can
demonstrate that behavior is something other than means of controlling
perceptual input, we would be a most attentive audience.

Bruce,

Why would you assume that a chaotic, dynamic system model would
contradict the idea that behavior is a means of controlling
perceptions? Any behavior that changes elements of the environment
must affect the perceptual inputs. it must be assumed that all
behavior is designed to change the environment (hence the perception)
in a beneficial way. No conflict arises here.

The point of contention is the method by which behavior is instigated
and and its effects are modelled.

Martin made the point in another post:

"One of the arguments Currently being used against PCT is that its
mechanism does not describe (or is not described by) current
understanding of neurological behaviour. Under my Occam's razor
argument, that's a valid complaint only if (a) the PCT model CLAIMS
to describe neurological data, and (2) it does so in a way that is
less accurate than is some competing mechanism."

Okay, the fact that the chaotic, dynamic system model takes into
consideration an understanding of how the brain works might be
irrelevant from the PCT point of view, but this doesn't negate the
second test for validity i.e.,providing a more accurate mechanism.

Let's take a popular element of PCT: the controlled variable. In PCT
these are always treated in isolation There is no consideration given
to the context in which they appear. Different contexts are assigned
to a vague "reorganization" process.

However, thinking in the real world places great emphasis on context.
The brain seems to be able to consider a controllable variable in
many different contexts and often resolves internal conflicts not by
changing the controllable variable, but also by changing the way the
variable is perceived. The PCT model doesn't provide any
visualization for how this might be achieved.

Let's look at how the chaotic, dynamic system model works.

It has been pointed out, in one of the papers I referenced, that it
was a great evolutionary advance when instinctive behavior could be
delayed for sufficient time to allow learned experiences to evaluate
a situation before action is put into effect. Then, if necessary,
responses can be cancelled or modified according to circumstance.

What happens during this time delay is that meanings are extracted
from the input from the senses. These meanings are used in various
combinations and with a variety of weightings to construct NOT A
SINGLE PERCEPTION but a series of perceptions. Each perception is
assembled individually from scratch and each perception triggers its
own unique variety of memories (learning) and emotions.

The brain rapidly cycles through the sequence of perceptions
manifested during this process, selecting the perception that appears
most appropriate (probably based upon some kind of summation of
emotions associated with the perceptions).

Brain imaging techniques seem to provide evidence of such a procedure
taking place. And, computer modelling has shown that the use of
attractors in this process would be an extremely fast and efficient
method to use.

It seems we are able to control the delay of behavior, sufficiently
long enough for enough different interpretations of the meanings to
be explored. Thus it can be assumed that control takes the form
cycling through many perceptions until an appropriate set of emotions
is registered. At this point the attractor responsible for this state
becomes dominant and the meanings and emotions associated with this
attractor take precedence over all others.

Control also seems to be related to how ingeniously the brain can
combine and give weight to meanings. This I would think would be a
function of experience and learning. It also appears that the brain
is able to simulate sensory inputs and their meanings and be able to
add these to those actually experienced.

The advantage of this model over PCT it that instead of viewing
controllable variables in isolation, they can be viewed in a variety
of different contexts. This accords more accurately with what seems
to happen in reality.

In the PCT model, behavior is seen as altering an external
controllable variable until the perception of the world matches some
internally create ideal. In the chaotic, dynamic system model, the
controllable variable is visualized in several different contexts and
behavior is instigated in light of one or other of the contexts. In
this model, control can also be exercised by changing the context in
which the controllable variable is viewed (i.e., choosing to act on
the basis of a different perception).

With the PCT model, context doesn't seem to come into it except
through some vague notion of reorganization. With the chaotic,
dynamic system model, it is necessary to consider not only the nature
of the controllable variables, but also the context in which you want
to view them - not easy to demonstrate with an elastic band.

The difference is that of one method is based upon techniques of auto
control, the other is based upon selection from a number of
alternatives. Current modern opinion seems to favor the latter.

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 [Marc Abrams (2004.05.13.0729)]

>[From Bruce Nevin (2004.05.12 11:40 EDT)]

>Peter, Marc, and others with expertise in biology, neurophysiology,
>neurochemistry, you chide us for our limited and outdated
knowledge of these fields. If you can demonstrate some other model in

which the...

You _completely_ misunderstand my position and feelings. I do not
'chide' anyone ever for believing in anything. I do criticize the total
insistence that the PCT model provided in B:CP is in fact the
physiological model as well.

It seems that you and MT have a fantasized version of what is going on
here on CSGnet.

What is PCT? Is it the contents of B:CP? Is it B:CP and the papers in
_Living Control Systems I & II_ or maybe either one, which at points
contradicts what is presented in B:CP. Is it the thoughts of Bill and
others over the years on CSGnet? Is it all of these things? If so, how
does one 'learn' what PCT is?

What is this forum about? I would think a clear definition of what PCT
is, and is not, might be a step in the right direction here. What
constitutes 'behavior'? What constitutes 'perception'? What constitutes
control?

Do cells 'behave'? Do cells perceive? I say they do, but does that make
it ok to talk about on CSGnet? I don't think so. Because the hierarchy
is no where to be seen in these controlled environments. Not yet anyway.

Is PCT, as MT suggests different than HPCT? If so in what ways?

I 'chide' the closed mindedness in this forum and the narrowness, myopia
and lack of creative ideas it produces, _not_ the fact that links to our
physiology have not yet been firmly established.

How do you propose these links are ever going to be established if you
don't move away from the current dogma and open up the discussion to new
ideas?

I'm far from being an 'expert' on physiology, but I'm working on it and
the view with control glasses on is spectacular. It's unfortunate that
discussions of this type are not viable on CSGnet yet because they are
certainly relevant to PCT. But I'm not giving up hope. I'll continue to
move in a direction I think useful and hopefully one day I'll meet a
kindred spirits here or elsewhere. I currently have a closed circle of
two very dear friends who share my views. For now that will suffice.

Marc

Something other than the tracking task is needed to provide data for the
control model. I believe it will come from biology and physiology.
Psychology and sociology have too many alternative explanations that
people are comfortable with.

Marc

Considering how often throughout history even intelligent people have
been proved to be wrong, it is amazing that there are still people who
are convinced that the only reason anyone could possibly say something
different from what they believe is stupidity or dishonesty.

Being smart is what keeps some people from being intelligent.

Thomas Sowell

From[Bill Williams 13 May 2004 7:40 AM CST]

From [Marc Abrams (2004.05.13.0729)]

> >[From Bruce Nevin (2004.05.12 11:40 EDT)]

You _completely_ misunderstand my position and feelings.

It is ok to be completely misunderstood by some of the people some of the
time, and it is ok to be somewhat misunderstood by all of the people some
of the time, and it is ok to be ... well you get the idea. However, if you
are completely misunderstood by all of the people all of the time-- that is
not OK.

Bill Williams

[From Bruce Nevin (2004.05.13 11:02 EDT)]

Peter Small (2004.05.13)--

Why would you assume that a chaotic, dynamic system model would
contradict the idea that behavior is a means of controlling
perceptions?

I wouldn't, and I didn't.

Any behavior that changes elements of the environment
must affect the perceptual inputs. it must be assumed that all
behavior is designed to change the environment (hence the perception)
in a beneficial way. No conflict arises here.

Perhaps there is only an appearance of conflict because this is an
extremely loose way of talking.

"Any behavior that changes elements of the environment must affect the
perceptual inputs" is obviously contrary to fact. We have effects all the
time on our environment that we do not perceive.

When you say that "all behavior is designed to change the environment
(hence the perception) in a beneficial way", what does the designing? What
determines whether the change is "beneficial" or not? As the perceived
state of the environment varies from "not beneficial" to "beneficial" as a
result of behavior, how does the mechanism generate just enough of just the
appropriate behavior, not too much, not too little, and neither irrelevant
nor ineffective? To take what seems to be a simple, concrete example, how
does your body maintain an erect posture? Or for another example, touch the
ring finger of your left hand to the word "posture" above. How do you do that?

The point of contention is the method by which behavior is instigated
and and its effects are modelled.

That's not a point of contention. It's a point of invitation. Specify your
model in sufficient detail and exactitude that a simulation can be
constructed and run. That is the methodology in this field.

Martin made the point in another post:

"One of the arguments Currently being used against PCT is that its
mechanism does not describe (or is not described by) current
understanding of neurological behaviour. Under my Occam's razor
argument, that's a valid complaint only if (a) the PCT model CLAIMS
to describe neurological data, and (2) it does so in a way that is
less accurate than is some competing mechanism."

PCT does not claim to describe neurological data. Various writings have
referred to neurological data in an attempt to demonstrate the plausibility
of the model.

Okay, the fact that the chaotic, dynamic system model takes into
consideration an understanding of how the brain works might be
irrelevant from the PCT point of view, but this doesn't negate the
second test for validity i.e.,providing a more accurate mechanism.

You have been invited to provide a more accurate mechanism. The mechanism,
however, must account for the phenomena of control. Not for the varieties
of behavior as commonly conceived. Conceptualizing behavior as an
identified something that is produced is a fundamental error of
reification. "Behavior" is variable all over the map. Controlled
perceptions are the stable entities. Indeed, to refer to behavior or
"behaviors" is to refer to perceptions controlled by the observer. The
epistemological difficulties that follow are obvious once you become aware
of this.

Let's take a popular element of PCT: the controlled variable.

It's not a "popular" element. If you don't understand what a controlled
variable is, you don't understand control, and you don't know what we're
talking about here.

In PCT these are always treated in isolation There is no consideration given
to the context in which they appear.

In a laboratory setting variables are isolated, as in any science. But to
say no consideration is given to varying context is quite wrong.

Different contexts are assigned to a vague "reorganization" process.

I don't understand this last sentence. Reorganization is described as a
last-ditch way of regaining control when other mechanisms for reducing
error fail. It has nothing to do with diversity of context within which a
given perceptual variable may be controlled.

However, thinking in the real world places great emphasis on context.
The brain seems to be able to consider a controllable variable in
many different contexts and often resolves internal conflicts not by
changing the controllable variable, but also by changing the way the
variable is perceived. The PCT model doesn't provide any
visualization for how this might be achieved.

You ought to get better acquainted before you make this kind of judgement.

Let's look at how the chaotic, dynamic system model works.

It has been pointed out, in one of the papers I referenced, that it
was a great evolutionary advance when instinctive behavior could be
delayed for sufficient time to allow learned experiences to evaluate
a situation before action is put into effect. Then, if necessary,
responses can be cancelled or modified according to circumstance.

Are these instinctive responses instances of control of input or not?

What happens during this time delay is that meanings are extracted
from the input from the senses. These meanings are used in various
combinations and with a variety of weightings to construct NOT A
SINGLE PERCEPTION but a series of perceptions. Each perception is
assembled individually from scratch and each perception triggers its
own unique variety of memories (learning) and emotions.

Is a meaning a perception? If not, what is it, and what is the difference?

The brain rapidly cycles through the sequence of perceptions
manifested during this process, selecting the perception that appears
most appropriate (probably based upon some kind of summation of
emotions associated with the perceptions).

Does it have to be a series or sequence? Could they not be constructed in
parallel?

Is an emotion a perception? If not, what is it, and what is the difference?
Is a summation of emotions a perception? If not, what is it, and what is
the difference?

When I asked you to point your finger at the word "posture" above, many
perceptions were available to you concurrently. Maybe you did it, maybe you
didn't. Choice, decision making, is not well specified in PCT. But once you
select which perceived state you prefer, the doing of it is a matter of
controlling your perception so that it reaches the preferred state.

I don't see any contradiction here.

It seems we are able to control the delay of behavior, sufficiently
long enough for enough different interpretations of the meanings to
be explored.

There is strong evidence that loop delay is longer at higher levels, as
predicted by HPCT.

Thus it can be assumed that control takes the form
cycling through many perceptions until an appropriate set of emotions
is registered. At this point the attractor responsible for this state
becomes dominant and the meanings and emotions associated with this
attractor take precedence over all others.

If you substitute "decision" or "choice" of what to control in place of
"control, there is no contradiction. The two descriptions are
complementary. At any moment you cannot control all of the huge number of
perceptions available, and the commitment of effectors to controlling some
perceptions makes them unavailable for controlling other perceptions. Once
you determine which variables to control, PCT accounts for the actual
behavior.

Control also seems to be related to how ingeniously the brain can
combine and give weight to meanings. This I would think would be a
function of experience and learning. It also appears that the brain
is able to simulate sensory inputs and their meanings and be able to
add these to those actually experienced.

PCT needs an account of learning to complement its account of behavior as
the control of perception. Various ideas have been advanced, none of them
sufficient in my opinion. Behaviorism, by the way, seems to me to be in
this complementary relation to PCT, as an investigation into learning,
especially reorganization processes under duress. It, too, says nothing
about behavior itself, only about establishment of preferences according to
which an organism controls its perceptions.

The advantage of this model over PCT it that instead of viewing
controllable variables in isolation, they can be viewed in a variety
of different contexts. This accords more accurately with what seems
to happen in reality.

You will have to explain why you believe that PCT precludes variables being
perceived and controlled in different contexts. I think that you have not
understood, but I could very well be misunderstanding you.

In the PCT model, behavior is seen as altering an external
controllable variable until the perception of the world matches some
internally create ideal. In the chaotic, dynamic system model, the
controllable variable is visualized in several different contexts and
behavior is instigated in light of one or other of the contexts. In
this model, control can also be exercised by changing the context in
which the controllable variable is viewed (i.e., choosing to act on
the basis of a different perception).

You seem to be saying that the perception of the relationship between the
visual image of my fingertip and the visual image of this word as I point
to it is different from the perception of the relationship between the
visual image of my fingertip and the visual image that picture frame across
the room because the context in which the fingertip is perceived is
different. But the relationship (superposition, or a finger's thickness
down and to the left, or circling around it, whatever the relationship may
be) is itself a perception that can be the same, or that may vary in the
same ways, regardless of the changes of context.

With the PCT model, context doesn't seem to come into it except
through some vague notion of reorganization.

Sorry, this just does not make any sense to me. Again, I have to conclude
that you have not understood, but I could very well be misunderstanding you.

With the chaotic,
dynamic system model, it is necessary to consider not only the nature
of the controllable variables, but also the context in which you want
to view them - not easy to demonstrate with an elastic band.

The difference is that of one method is based upon techniques of auto
control, the other is based upon selection from a number of
alternatives. Current modern opinion seems to favor the latter.

I think you are talking about selection of what to control. Once that
choice is made, you still have to account for how we act in just the right
way, and just enough, to "change the perception in a beneficial way". For
that, you need PCT.

         /Bruce Nevin

···

At 11:33 AM 5/13/2004 +0100, Peter Small wrote:

[From Bill Powers (2004.05.13.0814 MST)]

Peter Small (2004.05.13)--

Peter, you couldn't say most of the things you say about PCT if you had
actually read _Behavior: the control of perception_ and other instances of
my writings. You're basing what you say on hasty partial impressions,
skimming, and hearsay, which is a great way to risk making a fool of
yourself. Martin Taylor has been extremely patient in trying to help you
understand PCT, and you have responded with expressions of blind loyalty to
and faith in the current beliefs and practices in neuroscience and chaos
theory, topped with apparently endless erroneous statements about PCT,
Martin's knowlege, and anything else you wish to dismiss. I'm sure that
chaos theory has some useful things to say, but I am also sure that those
using it to analyze brain functions are being strongly biased by what they
already believe to be going on in the brain. Friedman found the olfactory
systems to be jam-packed with oscillators described by second-order
differential equations, a conclusion in which imagination and (I suspected)
instrumental artifacts played a far larger role than observation, at least
in the papers I read something like ten years ago. Again and again, in the
pages of Science and Nature, I see reports of discovering marvellous new
phenomena in the brain -- supported by brain-scan images with a few blobs
of activity here and there that could mean anything at all. What we are
seeing is theory run amok, dictating much of the data that are used to
support it, and suggesting overinterpretations that are about as objective
as descriptions of Rohrschach patterns.

The reason you don't see that sort of thing coming out of PCT is that I
refuse to do it. I prefer to keep claims modest and within the bounds of
what can be repeatedly and reliably demonstrated. Speculations are labeled
as such, and are always qualified appropriately, to the extent that I can
be conscientious enough. This puts me at a disadvantage in the marketplace
of ideas when others grab the spotlight by making exaggerated claims of
knowlege and accomplishment, but I am actually on much more secure ground
than they. I am in far less danger than they are of having to make excuses
and eat my words.

If you can explain behaviors like those we have studied in tracking
experiments, please do so -- not with theoretical imaginings, but with
convincing data and accurate analysis. If you can't do that, I see no
reason to put a lot of effort into understanding what you say.

Bill P.

From [Marc Abrams (2004.05.13.1128)]

[From Bruce Nevin (2004.05.13 11:02 EDT)]

That's not a point of contention. It's a point of invitation.
Specify your model in sufficient detail and exactitude that a
simulation can be constructed and run. That is the
methodology in this field.

Specifying, constructing and simulating a model is real easy. Getting
data that will make the model work (i.e. correct predictions) is
something entirely different. This has been the problem with the HPCT
hierarchy. Marken has shown that a hierarchy is a plausible idea. No one
has yet been able to get any data to justify the hierarchy as proposed
by Bill. In fact all attempts have failed.

Having a 'model' Bruce is meaningless if you don't have the data to plug
into it.

PCT does not claim to describe neurological data. Various
writings have referred to neurological data in an attempt to
demonstrate the plausibility of the model.

This is an interesting claim. Could you please tell me what Chapter's 7,
8, & 9 in B:CP are attempting to do?

What other writings are you referring too here?

You have been invited to provide a more accurate mechanism.

A more accurate mechanism to describe what? 'Behavior'? The behavior of
what? The organism? A cell? An organ system?

Conceptualizing behavior as an identified
something that is produced is a fundamental error of
reification.

I would think the HPCT hierarchy fits into the concept of reification as
well.

Peter Small:

>The brain rapidly cycles through the sequence of perceptions
>manifested during this process, selecting the perception
that appears
>most appropriate (probably based upon some kind of summation of
>emotions associated with the perceptions).

Sorry Peter, who or what 'selects' the 'appropriate' perception? This
'selection' process is still a deep mystery.
As is the construction of perceptions, regardless of what the hierarchy
might purport to represent.

Back to Bruce:

Is an emotion a perception? If not, what is it, and what is
the difference? Is a summation of emotions a perception? If
not, what is it, and what is the difference?

If you are seriously asking these questions then I think you need to
brush up on a bit on neuroscience. That is of course if your not
interested in redefining the entire spectrum of neuroscientific
research.

Please tell me what a 'perception' is in PCT speak? If by 'perception'
you mean a neural signal, then the answer is no. Emotions are largely
the result of chemical actions. Are chemical actions 'perceptions'? They
could be. Enzymes 'sense' the chemical environment. Do enzymes have
emotions? Maybe.

There is strong evidence that loop delay is longer at higher
levels, as predicted by HPCT.

You see, here is where I get bent out of shape.:slight_smile: The HPCT 'model' is
incapable of predicting anything. There is _no_ data for the higher
levels of the hierarchy so it can't very well produce data that doesn't
exist.

The HPCT hierarchy starts falling apart after a couple of 'levels'.

If you are talking _theoretically_ that is fine, but also meaningless.
Physics tells us we can _theoretically_ travel in time, stand a pencil
on it's point, and reconstruct an egg after it has been smashed. I would
put the plausibility of attainment of the HPCT hierarchy in this class
of events.

Once you determine which variables to control, PCT accounts
for the actual behavior.

And what accounts for the determination? Or what accounts for PCT?

PCT needs an account of learning to complement its account of
behavior as the control of perception.

When PCT comes up with a specific account for memory you'll have your
'learning' mechanism.

Marc

Considering how often throughout history even intelligent people have
been proved to be wrong, it is amazing that there are still people who
are convinced that the only reason anyone could possibly say something
different from what they believe is stupidity or dishonesty.

Being smart is what keeps some people from being intelligent.

Thomas Sowell

[From Bruce Nevin (2004.05.13 13:40 EDT)]

Marc Abrams (2004.05.13.0729)--

I do criticize the total
insistence that the PCT model provided in B:CP is in fact the
physiological model as well.

I'm not aware of anyone insisting on that. I've been asserting that it is
not, and certainly that, at least, makes the insistence less than total.

If you could descend from irate hyperbole to sober reportage, it would be
helpful to your cause. Give us an alternative physiological model to work
with. Don't just bitch about what a bad job Bill has done.

It seems that you and MT have a fantasized version of what is going on
here on CSGnet.

Well, OK. I do the best I can within the scope of my perceptions.

What is PCT? Is it the contents of B:CP? Is it B:CP and the papers in
_Living Control Systems I & II_ or maybe either one, which at points
contradicts what is presented in B:CP. Is it the thoughts of Bill and
others over the years on CSGnet? Is it all of these things?

It's a developing field and methodology of science, a field that is still
in its infancy. Inconsistencies between early writings and later writings,
and between research reports and philosophical speculations, are to be
expected.

If so, how does one 'learn' what PCT is?

If your aim is to pin it down to a fixed target that you can demolish,
you've got a tough task, not least because it will be damned hard to
demolish. If your aim is to participate in developing and extending it,
then all of the above are excellent resources for starters, but the best
way is to create a simulation of specific behavior, and, as a ramp-up to
that, to analyze carefully some of the better simulations that presently exist.

What is this forum about? I would think a clear definition of what PCT
is, and is not, might be a step in the right direction here. What
constitutes 'behavior'? What constitutes 'perception'? What constitutes
control?

I would refer people to http://www.ed.uiuc.edu/csg/tutorials.html. There
were efforts to develop a "Devil's Dictionary" some years ago, reviewing
common misconceptions and stumbling blocks. I think that should be buffed
up and posted there as well.

Do cells 'behave'? Do cells perceive? I say they do, but does that make
it ok to talk about on CSGnet?

Yes, yes, and I and others have talked about them.

I don't think so. Because the hierarchy is no where to be seen in these
controlled environments. Not yet anyway.

I don't understand your problem here. Either cellular control is organized
hierarchically or it is not. That's something to be found out. If you have
found out something about it, no one is stopping you from reporting that here.

Is PCT, as MT suggests different than HPCT?

Yes

If so in what ways?

In the ways that he and others have specified. HPCT understood as 11 levels
of perception each dwelling in an environment of the perceptions that are
controlled by the level below, down to sensors of the physical environment,
is an idealization, as well as being an extremely useful metaphor for
clinicians, for educators, for epistemologists, and so on. As Martin
pointed out, successful simulations like the little man don't accord neatly
with this idealization, nor do the data of the method of levels, at least
not if we take literally the metaphoric use of the term "levels" there. But
that said, I remember seeing citations of a lot of evidence for a
perceptual hierarchy in the conventional psych literature, and there are
quite a few solid demonstrations of hierarchical control on a small scale
(not the entire hierarchy). The DynaNeck demo may be an example, as Rick
has suggested.

I 'chide' the closed mindedness in this forum and the narrowness, myopia
and lack of creative ideas it produces, _not_ the fact that links to our
physiology have not yet been firmly established.

Nobody is stopping you from contributing something new. But no one can do
it for you.

         /Bruce Nevin

···

At 08:00 AM 5/13/2004 -0400, Marc Abrams wrote:

From [Marc Abrams (2004.05.13.1353)]

[From Bruce Nevin (2004.05.13 13:40 EDT)]

If you could descend from irate hyperbole to sober reportage, it would

be helpful to your cause. Give us an alternative

physiological model to work with. Don't just bitch about what a bad

job Bill has done.

This unfortunately is the CSGnet answer to my many questions. But Bruce,
you're absolutely off-base on a number of issues here. First, if I
thought Bill's work was horse-bleep I wouldn't be hanging around CSGnet.

Second, what 'irate hyperbole' are you referring too? The claim you made
about PCT not being made isomorphic too our neural systems and my
question about what Chapters 7-9 in B:CP represented that you failed to
answer in this post. So instead you turn too a personal attack and
called my questions 'irate hyperbole'. Give me a break. I expect more
from you then this type of stuff.

It's a developing field and methodology of science, a field
that is still in its infancy. Inconsistencies between early
writings and later writings, and between research reports and
philosophical speculations, are to be expected.

This is reasonable, but how does a person like Peter Small & others get
to 'learn' PCT? Who's version or interpretation is the valid one? If
it's Bill's, than how does someone know from day to day what the current
thinking is? Is Bill the final arbiter on what PCT is? What if the facts
provide a different picture (like memory, is it up to Bill to 'change'
PCT? It has been demonstrated through experimentation that our motor
control is not continuous and is chaotic. How does this affect the PCT
model?

I brought it up, why isn't is being discussed or commented on? I'll tell
you why. Because Bill isn't interested in anything that might challenge
his ideas.

I'm not aware of anyone insisting on that. I've been
asserting that it is not, and certainly that, at least, makes
the insistence less than total.

I'm not suggesting the insistence is total. I'm suggesting the
insistence is with one very influential person on CSGnet, Bill Powers

If so, how does one 'learn' what PCT is?

If your aim is to pin it down to a fixed target that you can
demolish, you've got a tough task, not least because it will
be damned hard to demolish. If your aim is to participate in
developing and extending it, then all of the above are
excellent resources for starters, but the best way is to
create a simulation of specific behavior, and, as a ramp-up
to that, to analyze carefully some of the better simulations
that presently exist.

You again, misrepresent my question. See my questions above. I asked
point blank, given the diversity of material and the change over the
years how does any new comer to CSGnet 'learn' PCT.

Creating a simulation of behavior is not a difficult task. It's getting
the data to make the model either descriptive of data you already have
or to produce data (prediction) that proves useful in duplicating the
behavior in the future. I can make, and many others have made, very fine
working SD models of human behavior, none of which utilize PCT but all
of which utilize both negative and positive feedback.

Ask Bill, about his frustrating meeting with Jay Forrester.

I am _firmly_ in Bill's camp here. The SD people are misguided and I'm
now in a joint effort with a well respected System Dynamist in trying to
correct this. We plan on presenting a paper in Spain at the next SD
conference (actually he will, I can't travel) and we are hopefully going
to publish a paper in January. He is _extremely_ enthusiastic and we are
both confident in what we are doing.

I would refer people to
http://www.ed.uiuc.edu/csg/tutorials.html. There were efforts
to develop a "Devil's Dictionary" some years ago, reviewing
common misconceptions and stumbling blocks. I think that
should be buffed up and posted there as well.

Do you remember my failed attempt at a PCT lexicon here on CSGnet. What
ever happened to Bjorn's attempt at it? I've done a lot more than just
bitch and moan. This is nice, but doesn't answer the need to have
something in one place that someone doesn't have to spend a year
researching simply to find out what one should know.

>Do cells 'behave'? Do cells perceive? I say they do, but
does that make
>it ok to talk about on CSGnet?

Yes, yes, and I and others have talked about them.

Really? Please point me too the archives. I've only been on CSGnet since
1994, so maybe these discussions happened before my arrival. _Principles
of Biological Control_ by David Horribin, 1970 MTP press provides a fine
description and many fine examples of biological control. I had the
pleasure of speaking with him before he passed away last year. He wrote
the book to make biological researchers aware of the phenomena of
control. He could not follow it up because he could not get any funding
for it. Sound familiar. He always loved the concept of control and
utilized it in his eventual research into AIDS and Cancer. His control
model would be easily understood by PCT'er's but is not the same as PCT.

I understand I might be considered heretical here but there are others
besides Horribin who have come up with behavioral control models. I will
provide the literature to anyone interested in seeing it.

My point here is _NOT_ to discredit PCT. My point is to show that there
is always more than one way to skin a cat and maybe, just maybe, PCT
could benefit from the work done by others. But we will never know this
on CSGnet, because this type of discussion would be viewed, and has been
viewed as a challenge to Bill's credibility. As a scientist this is
regrettable, as a person it's understandable.

I don't understand your problem here. Either cellular control
is organized hierarchically or it is not. That's something to
be found out. If you have found out something about it, no
one is stopping you from reporting that here.

I have 'reported' it, but no one seems interested here in microbiology.
Cellular control to date has found no hierarchy involved. Control may
very well be represented by different mechanisms at different scales of
observation. Empirical research will answer this question.

In the ways that he and others have specified. HPCT
understood as 11 levels of perception each dwelling in an
environment of the perceptions that are controlled by the...

And what is PCT?

Marc

Considering how often throughout history even intelligent people have
been proved to be wrong, it is amazing that there are still people who
are convinced that the only reason anyone could possibly say something
different from what they believe is stupidity or dishonesty.

Being smart is what keeps some people from being intelligent.

Thomas Sowell

···

level below, down to sensors of the physical environment, is
an idealization, as well as being an extremely useful
metaphor for clinicians, for educators, for epistemologists,
and so on. As Martin pointed out, successful simulations like
the little man don't accord neatly with this idealization,
nor do the data of the method of levels, at least not if we
take literally the metaphoric use of the term "levels" there.
But that said, I remember seeing citations of a lot of
evidence for a perceptual hierarchy in the conventional psych
literature, and there are quite a few solid demonstrations of
hierarchical control on a small scale (not the entire
hierarchy). The DynaNeck demo may be an example, as Rick has
suggested.

>I 'chide' the closed mindedness in this forum and the narrowness,
>myopia and lack of creative ideas it produces, _not_ the fact that
>links to our physiology have not yet been firmly established.

Nobody is stopping you from contributing something new. But
no one can do it for you.

         /Bruce Nevin

[From Bruce Nevin (2004.05.13 15:00 EDT)]

Marc Abrams (2004.05.13.1128) --

Specifying, constructing and simulating a model is real easy.

I'm looking forward to seeing yours.

Getting
data that will make the model work (i.e. correct predictions) is
something entirely different. This has been the problem with the HPCT
hierarchy. Marken has shown that a hierarchy is a plausible idea. No one
has yet been able to get any data to justify the hierarchy as proposed
by Bill. In fact all attempts have failed.

Having a 'model' Bruce is meaningless if you don't have the data to plug
into it.

Yes, part of the methodology is to obtain data about the state of a
supposed controlled variable and about all of the observable influences
that, in themselves, would change that state if they were not opposed by
other influences that are present. That is what the simulation must replicate.

Other data may be relevant. For example, detailed data about arm movement
was relevant for Bill's arm movement model, not because the movements
themselves are controlled variables, but to show that the several control
loops with their several degrees of freedom resulted in the same
non-obvious trajectory.

> PCT does not claim to describe neurological data. Various
> writings have referred to neurological data in an attempt to
> demonstrate the plausibility of the model.

This is an interesting claim. Could you please tell me what Chapter's 7,
8, & 9 in B:CP are attempting to do?

"Others who know more about behavior and many other subjects are the ones
to put the content in. Where I have tried to make the form more
comprehensible by suggesting content I have stepped over the bounds of my
knowledge and have probably made mistakes. I trust that they can be
corrected without obscuring what is most important about this model -- and
that is this: Behavior is the process by which organisms control their
input sensory data. For human beings, behavior is the control of
perception. That is what is important about the model -- that, and all that
is implied by it." B:CP p. xi.

What other writings are you referring too here?

I had nothing specific in mind. I remember seeing references in various
places, but I'm not going to go look them up. There are a good number of
such references in the CSGnet archive. In every case, as I recall, these
references were in a secondary and supportive role, as suggesting the
biological plausibility of the model, and hedged with caveats about the
limitations of our understanding.

> You have been invited to provide a more accurate mechanism.

A more accurate mechanism to describe what? 'Behavior'? The behavior of
what? The organism? A cell? An organ system?

You pick, Marc. What are you interested in? Do you think it is controlling
its sensed input? Then it's fair game. Do you think that within the
organism the control loop comprises attractor basins instead of input
functions, comparators, output functions, and neural signals? Then specify
in those terms the mechanisms for implementing the portion of the control
loop that resides within the organism. Don't just tell us we're wrong. Show
us what's right.

>Conceptualizing behavior as an identified
> something that is produced is a fundamental error of
> reification.

I would think the HPCT hierarchy fits into the concept of reification as
well.

Could be. We have good data about behavior and control. The data about the
perceptual hierarchy are less solid. This is immaterial to the point.

Back to Bruce:

> Is an emotion a perception? If not, what is it, and what is
> the difference? Is a summation of emotions a perception? If
> not, what is it, and what is the difference?

Emotions are largely
the result of chemical actions. Are chemical actions 'perceptions'? They
could be. Enzymes 'sense' the chemical environment. Do enzymes have
emotions? Maybe.

OK, they are largely the result of chemical actions. And the increase in
the temperature of the air in this house today has been largely a result of
radiation from the sun. The ambient temperature is not the radiation. the
chemical actions are not the emotions. These changes in the body can all be
sensed. If you don't perceive it, or can't perceive it, is it an emotion?
If so, you are using the term in an unusual way. Everywhere else, the word
is limited to things that people can perceive. Is it an emotion before you
sense it? I am quite willing to accept that it is, and to adopt (here) a
modification of the usual definition of "emotion". I'm trying to find out
if that is what is required in order to talk with you.

> There is strong evidence that loop delay is longer at higher
> levels, as predicted by HPCT.

You see, here is where I get bent out of shape.:slight_smile:

Then omit the last phrase. There is strong evidence that loop delay is
longer at higher levels of the perceptual hierarchy.

> Once you determine which variables to control, PCT accounts
> for the actual behavior.

And what accounts for the determination? Or what accounts for PCT?

I can't make any sense of this.

> PCT needs an account of learning to complement its account of
> behavior as the control of perception.

When PCT comes up with a specific account for memory you'll have your
'learning' mechanism.

OK. Is that something you're working on? That would be an excellent
contribution to PCT!

         /Bruce Nevin

···

At 12:51 PM 5/13/2004 -0400, Marc Abrams wrote: