Re: The switching or the mixing function
(memory).
[Martin Taylor 2006.03.26.10.56]
[From Bjorn
Simonsen (2006.03.26,13:35 EUST)]
I am still working
with Martin’s World Model. Now I have lent P.E.Wellstead
:Self.Yuning Systems.
Today I ask myself
the question why Martin and other needed a World Model. From
Martin’s mails I imagine his answer is something like this: " We
need a model that predicts an imagined result in such a way that we
can perceive it without any disturbances from the World out
there".
Not at all. It has nothing to do with disturbances. It has to do
with the likely effects of choosing different courses of action. Do I
take the bike or the car? My World model tells me that if I take the
bike I must start earlier or I’ll be late, that if I take the car I
won’t get the benefit of exercise or fresh air or the feel of the
scenery, that if I take the bike I can go along the path along the
river and enjoy the quiet whereas if I take the car I’ll be stuck in
smelly traffic on the highway, etc. etc.
The probable perceived consequences (perceptions resulting from
actions) don’t have much to do with disturbances, because the World
Model at any one level has been “self-tuned” by situations
in which control at the lower levels is pretty effective. Of course,
when a behavious is executed in the real world (one actually has taken
the bike or the car), the resulting perceptions are subject to
disturbances which may be outside the range of control. Imagination
doesn’t always correspond to reality!
Basis for my
thoughts.
- We use and
understand the PCT loop as a reflex loop, as a composite loop
representing many loops and as a human composite loop representing
general human behavior. The loop is a model we use to explain how
human control different variables.
OK.
- In a human brain
there are a lot of loops, thousands and maybe
thousands.
I have asked myself
the question: “How is the state of the loops we don’t control?”
And I have got answers like: “They are always
functioning”.
Thogh most of them are broken, in the sense that they don’t
generate action in the real world, and so are not, at that moment,
functioning as loops. So far as we know, though, there are no such
limitations on the number of loops acting only through imagination.
Nor, put more broadly, is there any known limitation on the number of
loops that exist only within the brain, not using the peripheral
nervous system or the musculature.
The words “loops we don’t control” are puzzling. In
what sense do we ever “control” a loop? Do you mean that we
are changing its parameters or its structure in order to make it more
efficient (i.e. that we are reorganizing in a way that concerns this
particular loop)? I haven’t seen reorganization characterized in this
way, though it’s certainly a feasible proposition. How do you see it
operating – specifically, what property of the loop would the system
that controls it be able to perceive in order to control the
loop?
- Since most of
the feedback loops are negative, the normal state, if no disturbances
are affecting, of the loops will have a zero error
state.
In PCT, a control loop never has zero error in the presence of
disturbance, even when a single elementary control system (ECS) is
considered by itself.
If there is more than one ECS in the complex system, the
possibility of conflict exists. Conflict means it is not possible that
all the ECSs have zero error. If there are more active ECSs than
degrees of freedom at the bottleneck in the common pathways (in
zoological systems including humans, that’s at the muscular output),
then conflict is certain to exist and zero error is impossible to
achieve, even in the absence of disturbance.
The switches
don’t exist as a mixing function. We don’t need a model of a
switching or mixing function. We neither need a World Model. The
switches and the mixing functions are our
awareness.
Strange. If I understand you correctly, you say we don’t need
these things (presumably you mean as theoretical constructs; whether
we need them in reality is something we can’t know – yet). Then you
go on to say that althoguh we don’t need them, we have them, and
define where we have them. You aren’t actually contradicting yourself,
but you are saying we use stuff we don’t need. It may be so, but it’s
not the Accam’s razor solution. If we don’t need them, we don’t need
them, period. And if we don’t need them, why should we then have
them?
When we are
aware of perceptions representing disturbances and feedback signals
from the world out there, we are in control mode.
??
Surely we are in control mode when our actions are countering the
perceptual effects of external disturbances. What does this have to do
with being aware of the controlled perceptions? Typically, it’s the
perceptions that we are least aware of that are under the best
control, no?
When we are
aware of perceptions that are not affected of disturbances, we are in
remembering mode.
?? I don’t really understand what you mean here. I can certainly
remember lots of things, and those memories may not be affected by
current disturbances. But there are other perceptions not affected by
current disturbances. They may be composed of remembered elements but
are not themselves memories. If you are capable of mental imagery (not
everyone is), I bet you can imagine a green elephant with star-shaped
purple patches walking in your hometown main street. And I’ll bet you
never saw such a thing in real life. And I’ll further suggest that
this perception is not a memory, nor is it affected by current
disturbances.
In the same way as
we are able to control different variables in real time control we are
also able to control different variables in remembering mode. Then we
control our perceptions in the imagination mode.
Yes. That’s where the World model comes in, to allow that to
happen. I just don’t follow your use of the word “Then”. I
think imagination uses memory, but is more than memory.
When we are
conscious but don’t control neither in control mode nor in
remembering/imagination mode we are in passive observation
mode.
Yes, but it doesn’t need to have anything to do with being
conscious. The same thing presumably happens with perceptions of which
we are not conscious. That’s harder to prove directly, since changes
in uncontrolled perceptions don’t evoke externally detectable actions,
but it would be required in most versions of PCT, including the
classic hierarchy.
Subjectively, you can observe it. I’m sure you have been in a
room and have heard a fan or some other noise switch off, even though
you had not been consciously hearing it when it was steady.
Retrospectively, you heard the last few seconds before the switch-off.
Surely that was a sound you were passively, but not consciously,
observing all along. It’s most unlikely that your passive observation
started three seconds before the fan switched off. Hard to prove
experimentally, though.
A reason why our
perceptions in remembering or in imagination mode are more vague is a
result of no effect from disturbances.
Why?
When
disturbances affect our sensory cells, the perceptual signals are
different from the zero error state. And our ability to perceive
something is affected by the degree the perceptions are different from
perceptions in zero error state.
Why?
Both these assertions are simply that – assertions. Why would
anyone expect that our ability to perceive something is affected by
the level of error? That assertion is certainly in direct
contradiction to the normal assumptions of PCT. I thnk such an
assertion needs some kind of evidence or demonstration of theoretical
superiority (by that, I mean that it has to be shown to predict data
better than normal PCT does).
Subjectively, it is true that those things of which we tend to be
conscious are those that we have difficulty controlling, together with
those from which or to which we may be soon switching control when we
are time-multiplexing. But that’s just the conscious representation.
What this suggests to me is that consciousness is related to the
switching function. That’s not the same as your thesis that there is
no switching function, its place being taken by consciouslness.
Has anybody
anything to say?
No
Martin