Memory

[_Martin_Taylor]

[Martin Taylor 960124 11:30]

Bill Powers (960123.1345 MST)

+Bruce Abbott (960124.1055 EST)

I think we can agree that one of the problems a model has to deal with
is the way organisms seem to remember what to do to achieve a particular
end. But that isn't the only way to describe the problem. "Remembering
what to do" can equally well be described as "acquiring a connection
between a higher-order error and a lower-order reference signal." This
doesn't necessarily involve memory, unless you want to include the
establishment of a neural output connection in the definition of memory
(which I don't want to do).

+This is wise; the changes in neural connectivity thought to underly memory
+are probably only a subset of the ways in which a neural output connection
+can be established. For example, connections can be effectively "changed"
+on the fly by neural inhibitory elements in much the same way that a signal
+applied to one leg of a digital "and" gate can determine whether the signal
+applied to the other leg is "passed through" or not. We probably wouldn't
+want to label such dynamic changes "memory" unless the altered conductivity
+arose from more permanent changes in neural structure, and even then a
+"memory" would not reside in any individual neuron.

Insofar as we have no idea about what we are not aware of, I go along with
not wanting to identify changes in neural connectivity with ALL of memory,
but how can one not say that a change in connectivity is a part of memory?

Apart from recurrent circuits and possibly analogous biochemical circuits,
I'm not aware of changes other than neural connectivity that are easily
identified with memory. For that reason, any time someone mentions "memory"
I tend to translate it into "changes in neural connectivity" unless there
is reason in that specific case to treat it otherwise.

Bruce's objection, that the dynamic behaviour of a particular connection
varies with the ongoing data, seems less of an objection than a description
of how we remember something at one moment that we could not remember a
little earlier. In other words, it supports the idea of memory as
connectivity, rather than providing an objection.

But then I have no idea of what I am not aware of, and can now remember
only what I remember now.

Martin

[BJORN_SIMONSEN]

[From
Bjorn Simonsen (2004.0X.XX.XXXX
EuST)]

I will bring forward some knowledge about Memory. I
think it fits in with PCT but here is emphasized on something new for me. I
refer to a Norwegian radio program yesterday (Clive Bramham).

When we control perceptions, we often learn new
aspects of knowledge that is already in our brain as memory. I think this is an
example where Learning is Reorganization. This is my wording so far. I’ll give
an account of what I heard in brief outline what I think we know and boldface
what is new for me.

1.  In a set
   

of many functioning ECUs (Elementary Control Units), we can follow
impulses of frequencies. They pass the gaps = Synopsis on their way to
next nerve cell.

1.  The Synopsis
   

grows larger when an ECU is functioning.

1.  There
   

is established more synopsises when an ECU is functioning.

1. There is a BDNF protein (** ‘‘Brain-derived neurotrophic factor’’) and he has
   studied this factor in Hippocampus. In some other papers he emphasize that
   there are no specific memory centres in the brain. This fits in with PCT.**

2. ** When one nerve
   cell feed another nerve cell with impulses, the other cell becomes a
   better recipient after 20 milliseconds. The increasing of this activity
   can be stable for many weeks. In this way memory establishes. Being a
   better recipient is a process called LTP (long time potential).**

3.  During LPT, the BDNF reinforce the synopsis and mRNA contributes
   

to the production of protein which stable memory.

1.  The receiving cell doesn’t respond very well on old rats.
   

(It is bad to be old (my words))

1.  They think antidepressants work through BDNF. Stress reduses
   

the production of BDNF and activity stimulate the production of BDNF.

1. When rats sleep the BDNF not produced.

2.  Memory is established when we are awake. They think
   

memory is consolidated under REM sleep, but I don’t understand how.

I am going to France tomorrow. I will be back in 10
days and then I’ll continue with a memory mail.

Bjorn

PS. I send a copy of this mail to Clive Bramham
(Professor, dr. Philol), University in Bergen. I don’t think is a PCR-er.

···

[Bill_Powers1]

1. In a set of many functioning ECUs (Elementary Control Units), we can
   follow impulses of frequencies. They pass the gaps = Synopsis on their
   way to next nerve cell.
   [From Bill Powers (2004.05.06.0630 MST)]

Bjorn Simonsen (2004.0X.XX.XXXX
EuST)–

“Synapses” (the plural of synapse, a connection from one
nerve cell to another) is the term. “Synopsis” means a brief
summary of a longer statement or story.

Impulses, as far as I know, are not individually relayed from input to
output. Instead, the neurotransmitters from all incoming signals diffuse
from the dcentrites into the cell body, where they influence the voltage
at the axon hillock, which in turn determines the average rate of firing
of the cell. Only in certain special cells (“electrical” cells)
is there any one-to-one correspondence of input to output
impulses.

1. The Synopsis grows larger when an ECU is
   functioning.
   I don’t know what it means to say that a synapse “grows
   larger” Does the synaptic button at the end of the incoming axon
   increase its size? Does one impulse release more neurotransmitters than
   before? Do the receptive sites on the receiving dendrite grow in
   size?

1. There is established more synopsises when
   an ECU is functioning.
   If the number of synapses increases, the same input neural signal
   will produce a higher rate of firing of the receiving cell body. Does
   this mean that when an ECU is functioning, its input weights are
   continuously increasing? Is that
   good?

1. When one nerve cell feed another nerve
   cell with impulses, the other cell becomes a better recipient after 20
   milliseconds. The increasing of this activity can be stable for many
   weeks. In this way memory establishes. Being a better recipient is a
   process called LTP (long time potential).
   I fail to see how this explains memory. A perceptual input function,
   as I understand it, computes some function of its input signals to
   produce its output signal. The nature of the function is determined by
   the positive and negative weightings given to incoming signals, as well
   as other more complex computing functions such as differentiation and
   integration , multiplication, and thresholding. Simply increasing the
   weighting on one signal (or all signals) would not seem to accomplish
   much, and could destroy the perception being computed. I don’t think that
   increasing the sensitivity of an input over 20 milliseconds means
   anything but a time-function in the computation.

1. During LPT, the BDNF reinforce the
   synopsis and mRNA contributes to the production of protein which stable
   memory.
   I really don’t see how increasing the sensitivity of a synapse to
   incoming signals can explain the fact that I remember eating a doughnut
   10 minutes ago. There is a huge gap between a description of some
   localized electrochemistry and the phenomenon we experience as
   remembering things.
2. The receiving cell doesn’t respond very
   well on old rats. (It is bad to be old (my words))
3. 1. When rats sleep the BDNF not produced.
      They think antidepressants work through
      BDNF. Stress reduses the production of BDNF and activity stimulate the
      production of BDNF.
4. Memory is established when we are awake.
   They think memory is consolidated under REM sleep, but I don’t understand
   how.
   Neither do they. Perhaps you can tell that I don’t like this sort of
   imaginative overinterpretation, which grossly exaggerates the amount of
   understanding we get from some relatively simple observations.

Best,

Bill P.

[BJORN_SIMONSEN]

[From
Bjorn Simonsen YG(2004.05.06.18:10 EuST)]

[From
Bill Powers (2004.05.06.0630 MST)]

In
a set of many functioning ECUs (Elementary Control Units), we can follow

impulses of
frequencies. They pass the gaps = Synopsis on their way to next nerve cell.

“Synapses” (the plural of synapse, a
connection from one nerve cell to another) is the term.

“Synopsis” means a brief summary of a
longer statement or story.

Thanks
for the Synopsis. I appreciate learning anything. I’ll read your mail again
when I am well back again. Then I also will book his thesis for the doctorate.
I guess I haven’t expressed myself to his advantage. Have in mind Synapses-Synopsis.

You’ll
hear from me.

bjorn

[Marc_Abrams2]

from [ Marc Abrams (990707.1154) ]

Bill, have any of your thoughts and/or ideas changed at all with regard to
Chap. 15 ( Memory ) in B:CP? If so, Is there more recent PCT lit you can
refer me to. If not I have a few questions.

1) Am I correct in assuming that memory exists within each control loop?
2) _Each_ of the 4 modes you explained ( Passive Observation, Automatic,
Imagination, Control ), could be in _any_ one control loop?
3) Do you think it plausible that various memory patterns among and between
levels might explain, why we do and don't remember certain things?
4) Do you have any ideas about what kind of methods and data you would need
to test this conjecture?

Thanks,

Marc

[Bill_Powers1]

[From Bill Powers (990708.1912 MDT)]

Marc Abrams (990707.1154)--

1) Am I correct in assuming that memory exists within each control loop?

I proposed that, but there hasn't been any research so it's still just a
proposal.

2) _Each_ of the 4 modes you explained ( Passive Observation, Automatic,
Imagination, Control ), could be in _any_ one control loop?

Yes, if all control systems are organized like that diagram with the
switches. Again, this is just a proposal. Do you have any evidence one way
or the other?

3) Do you think it plausible that various memory patterns among and between
levels might explain, why we do and don't remember certain things?

Don't know. It's plausible, but is it true?

4) Do you have any ideas about what kind of methods and data you would need
to test this conjecture?

I really haven't done ANY research about this. I once thought I would, but
things haven't worked out that way.

Best,

Bill P.

[Marc_Abrams2]

from [ Marc Abrams (990708.2319) ]

Bill, I am simply trying to understand what the current thinking is on
memory, within the PCT model.
You say on pgs 216-217 of B:CP, ..."I will employ this principle ( i.e.
"Memory, in order to be useful, must be under control of some orderly
addressing mechanism" ) to give the hierarchy of control systems a new
property which accounts not only for perceptual remembering, but for the
ability of organisms to reproduce past perceptual situations through
actions. Once it is noticed that the model lacks this ability, one sees how
essential it is.
    As it stands now, the model contains nothing to evoke memories; no
source of addresses for the memory box of figure 15.1 ( pg. 209 ).
Furthermore, there is not even a way for the system to re-create a specific
past perceptual signal, for reference signals are generated independently of
perceptual signals and only by chance would demand recreation of a
perceptual situation that had actually occurred before. There are many
situations in which past perceptual situations _are_ _ ( Bill. italicized
the word ) re-created, quite specifically and precisely".

You also say on pg. 206 ..."Here ( in talking about how memory info is
retrieved and accessed ) behavioral experiments and direct experience may be
as valuable as biochemical investigations in telling us what capabilities to
look for."

Pretty powerful statements.

In my ongoing pursuit of SD_MOL, the question of "Is it live, or is it
Memorex" ( borrowed from an old Memorex TV ad, Memorex produces magnetic
audio and video tapes ) Looms larger and larger. What effect does memory
have on current perceptions?, Reference signals? These questions become
important when "looking" at past experience.

In my 5 years ( or is it 6 ) on CSGnet I have not seen any discussion
specifically on the proposed modes. ( pg 220 B:CP ) of memory. Does anyone
know differently and can point me to some postings in the archives?.

I think this is an important area. Does anyone else?

[From Bill Powers (990708.1912 MDT)]

>2) _Each_ of the 4 modes you explained ( Passive Observation, Automatic,
>Imagination, Control ), could be in _any_ one control loop?

Yes, if all control systems are organized like that diagram with the
switches. Again, this is just a proposal. Do you have any evidence one way
or the other?

No, but we do need a starting point. I think that provides a good one. I
just wanted to make sure that your thinking on it has not been revised.

>3) Do you think it plausible that various memory patterns among and

between

>levels might explain, why we do and don't remember certain things?

Don't know. It's plausible, but is it true?

Don't know, anyone interested in trying to find out?

>4) Do you have any ideas about what kind of methods and data you would

need

>to test this conjecture?

I really haven't done ANY research about this. I once thought I would, but
things haven't worked out that way.

That's OK, Shit happens. You can't be everything to everyone. I would hope
that you might get involved in a discussion if it attempted to look at the
possibilities.

Marc

[Bill_Powers1]

[From Bill Powers (990709.0409 MDT)]

Marc Abrams (990708.2319)--

Bill, I am simply trying to understand what the current thinking is on
memory, within the PCT model.
You say on pgs 216-217 of B:CP, ...

Right, and that's it as far as "current thinking" is concerned. To get any
father than those proposals, SOMEONE is going to have to do some
experimental research. The proposals explain certain phenomena, and if I
say so myself they do so rather economically. But that's only the first
step. The next step is to devise experimental conditions that challenge
this model, and find out if it will hold up under pressure. That hasn't
been done. You know why; I don't have to go through all that again.

In my ongoing pursuit of SD_MOL, the question of "Is it live, or is it
Memorex" ( borrowed from an old Memorex TV ad, Memorex produces magnetic
audio and video tapes ) Looms larger and larger. What effect does memory
have on current perceptions?, Reference signals? These questions become
important when "looking" at past experience.

In my 5 years ( or is it 6 ) on CSGnet I have not seen any discussion
specifically on the proposed modes. ( pg 220 B:CP ) of memory. Does anyone
know differently and can point me to some postings in the archives?.

There's been plenty of not-very-well-organized research on memory phenomena
by conventional psychologists, but none that I know of designed to test my
hypotheses. You may be hoping you can just look up the answers you need,
but I'm afraid that this field is at the point where you have to do the
work yourself if you want the answers badly enough.

Best,

Bill P.

[Marc_Abrams2]

from [ Marc Abrams (990709.0916)]

[From Bill Powers (990709.0409 MDT)]

Marc Abrams (990708.2319)--

>Bill, I am simply trying to understand what the current thinking is on
>memory, within the PCT model.
>You say on pgs 216-217 of B:CP, ...

Right, and that's it as far as "current thinking" is concerned. To get any
father than those proposals, SOMEONE is going to have to do some
experimental research. The proposals explain certain phenomena, and if I
say so myself they do so rather economically. But that's only the first
step. The next step is to devise experimental conditions that challenge
this model, and find out if it will hold up under pressure. That hasn't
been done. You know why; I don't have to go through all that again.

No question. It's kind of exciting. I think you provide a terrific anchor
and base to work from. A couple of things start taking on a slightly
different focus when your proposals are integrated into the model.
1) You can begin to see that the 4 modes ( i.e the 3 "memory" modes and the
"control mode" ) as you defined them, do not seem to be "used" in arbitary
patterns. I believe this can be looked at.

2) I can begin to appreciate how quickly our lower levels operate. It
memory ) also opens up the possibiltiy that the higher levels of the
hierarchy might be just as much of a function of our memories as "levels".
Something I can start to see in my SD-MOL's. mapping. This is something else
I think can be looked at.

3) I believe that this can in fact be modeled in Vensim. The interesting
thing about this is that you would have to make certain conjectures about
the levels ( PCT levels, that is ) and the properties they have. Not quite
up to that but I don't see that as an impossible or undoable task.

4) I think the MOL and SD-MOL could be terrific tools for data gathering for
this purpose.

There's been plenty of not-very-well-organized research on memory

phenomena

by conventional psychologists,

Useless.

but none that I know of designed to test my hypotheses.

That's what I wanted to know. Do I hear opportunity knocking?

You may be hoping you can just look up the answers you need,
but I'm afraid that this field is at the point where you have to do the
work yourself if you want the answers badly enough.

Yes that would have been nice. But I'll take what I got. I just find it hard
to believe that this stuff has been totally passed over, all these years. I
think it's one of the links that connect PCT to the "real" world and it's
"applied" use in it. This is gonna be fun. I think I found a niche for
myself.

Marc

[R_Marken]

[From Rick Marken (990709.0950)]

Marc Abrams (990709.0916)

I just find it hard to believe that this stuff [the PCT memory
model] has been totally passed over, all these years.

Do you find it easier to believe that the basic control of
perception model of behavior has been totally passed over
(by conventional psychologists) all these years (at least 40
of them by now)?

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[Jeff_Vancouver2]

[from Jeff Vancouver 990709.1447]

From [ Marc Abrams (990708.2319) ]
I think this is an important area. Does anyone else?

Yes. I think the lack of developing this aspect partially accounts for the
lack of conventional psychologists' attention to the theory. Most
psychologists have concerned themselves with phenomenon related to the
other modes (besides control), but they just do not realize it.

Sincerely,

Jeff

[Marc_Abrams2]

from [ Marc Abrams (990709.1438) ]

[From Rick Marken (990709.0950)]

Do you find it easier to believe that the basic control of
perception model of behavior has been totally passed over
(by conventional psychologists) all these years (at least 40
of them by now)?

Yes. They don't know any better. And Ironically if we better understood
Bill's proposals in Chap 15 we might have a better understanding as to _why_
it has turned out this way. We ( those on CSGnet ) should. Chap 15 becomes
extremely important, in my view, when you have some understanding of the
"basic" PCT model. PCT is not simply the "basic" control model. There are 3
other modes that were proposed by Bill that need to be examined, and are
just as important as the basic model. In fact, you would be hard put to
explain "real" behavior without them. Knowing and explaining PCT might be a
bit _easier_ if we had some understanding of what Bill was proposing in Chap
15. I think the work is _brilliant_ . Chap 15 is of no use if you do not
understand the "basic" model. But if you do? ( as we on CSG, or at least
part of us claim ).
I can say that this is history, at least with regard to one individual.
Enough ranting. I got a lot of work to do

Marc

[Marc_Abrams2]

from [ Marc Abrams (990709.1504) ]

[from Jeff Vancouver 990709.1447]

Yes. I think the lack of developing this aspect partially accounts for

the

lack of conventional psychologists' attention to the theory. Most
psychologists have concerned themselves with phenomenon related to the
other modes (besides control), but they just do not realize it.

Jeff, _All_ the modes are about control. _Nothing_ happens without it. The
key is in _how_ we aquire, maintain, and change _what_ we control. Have you
read Chap 15? If you have, re-read it. If you haven't or can't recall the
details, read it again. From a modeling perspective I think you will find it
_very_ interesting. The possibilities are huge.

Marc

[Isaac_Kurtzer1]

i.kurtzer (990709.1515)

[from Jeff Vancouver 990709.1447]

>From [ Marc Abrams (990708.2319) ]
>I think this is an important area. Does anyone else?
>

Yes. I think the lack of developing this aspect partially accounts for the
lack of conventional psychologists' attention to the theory. Most
psychologists have concerned themselves with phenomenon related to the
other modes (besides control), but they just do not realize it.

I respectfully disagree. I think that the lack of attention is because:
a. there are almost no PCT experiments
b. there is virtually no systematic attempt to explore PCT within a field--as
opposed to demonstrations of principle, which are wonderful but not cumulative
science.
c. to wax Rickishly: they don't know what control is, if it bit them on the
ass.

my answer, do B to the best of your ability within whatever stuff/thing/science
it is that you do.

i.

[Marc_Abrams2]

from [ Marc Abrams (990709.1552) ]

i.kurtzer (990709.1515) replying to
> [from Jeff Vancouver 990709.1447]

I respectfully disagree. I think that the lack of attention is because:
a. there are almost no PCT experiments
b. there is virtually no systematic attempt to explore PCT within a

field--as

opposed to demonstrations of principle, which are wonderful but not

cumulative

science.
c. to wax Rickishly: they don't know what control is, if it bit them on

the

ass.

my answer, do B to the best of your ability within whatever

stuff/thing/science

it is that you do.

Nice post. I agree with all of it. I especially like b. I think for those
interested in exploring PCT in applied Psych areas I think Chap 15 provides
some important aspects of PCT that have not been explored yet, and in my
opinion no less important then the "basic" model.

Marc

[Jeff_Vancouver2]

[from Jeff Vancouver 990709.1650]

···

At 03:09 PM 7/9/99 -0400, you wrote:

From [ Marc Abrams (990709.1504) ]

Jeff, _All_ the modes are about control. _Nothing_ happens without it. The
key is in _how_ we aquire, maintain, and change _what_ we control. Have you
read Chap 15? If you have, re-read it. If you haven't or can't recall the
details, read it again. From a modeling perspective I think you will find it
_very_ interesting. The possibilities are huge.

I agree, they are all about control, but only one mode is labeled control
(although I prefer, behaving for the very reason you misunderstood me).
The chapter is extremely important. It relates to much of the work I am
doing (theoretically anyway, empirically it is clearly more challenging).

Sincerely,

Jeff

[Marc_Abrams2]

from [ Marc Abrams (990709.1717) ]

[from Jeff Vancouver 990709.1650]

I agree, they are all about control, but only one mode is labeled control

Yes, at first It didn't sink in. I thought it was simply another "memory"
mode. Then I realized that Bill was actually talking about how _all_ 4
modes were needed to describe "behavior". Yet we seemed to have concentrated
on one only.

(although I prefer, behaving for the very reason you misunderstood me).
The chapter is extremely important. It relates to much of the work I am
doing (theoretically anyway, empirically it is clearly more challenging).

Can you see yourself ultimately modeling, with these different modes? Would
you like to talk about the possibilities? and some possible strategies for
doing so?

Marc

[R_Marken]

[From Rick Marken (990709.1510)]

Jeff Vancouver (990709.1655) --

I think Rick would not see a conventional psychologist studying
control unless it bit him in the ass (i.e., used the test for
the controlled variable).

How could _anyone_ study control without testing for a controlled
variable?

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates mailto: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[Marc_Abrams2]

from [ Marc Abrams (990711.1653) ]

[From Bruce Nevin (990711.1406)]

I am looking forward to seeing what you come up with. FWIW, your
conjectures make sense to me. But I'm sure that modelling and
experimentation will bring some surprises.

I think so to :-). I'm going into digest mode, so I can spend more time on
this, and less time on non productive net communications. It seems obvious
to me at this point that no one is either capable of or willing to engage
_me_ in constructive dialogue on Bill's proposal's in Chapter 15. Bruce, I'm
going to take some old advice from you, and come back when I have a
reasonable model to present.

Excuse me folks while I take the SD-MOL and some conjectures about memory
and it's effects on the control model and go have some fun :-). If you need
me, you know where to find me.

To all those attending, have a great conference.

Marc