Right decisions

[Richard_Marken]

[From Rick Marken (960226.1230)]

Stefan Balke (960226.13.40 MEZ) --

I mean the case when the object of desire is still unvisible or even
completly unknown...when there is no direct feedback about coming closer to
or getting farer from the intended result, because neighter the way my
actions are perceived from the relevant person nor the indended result
itself is well known.

There is always direct, visible feedback about the state of a controlled
perceptual variable. Direct visible feeback is just the perception of the
state of the controlled variable.

It is important to remember that what we control (or try to control) are
perceptual _variables_; aspects of our experience that can change states over
time. For example, we don't control a "car" per se; rather, we control
variable perceptual aspects of a car; its position on the road, its speed,
its color, etc. The controlled perceptual variable in your example above is a
_relationship_ variable; it might be called the "degree to which I have a
loving relationship with a suitable girlfriend"; we can call this variable
"girlfriend" for short.

When you start controlling for a particular (reference) state of the
girlfriend perception (let's call that reference state "Michelle Pfieffer")
you are probably perceiving something quite different than the reference
state of the perception - - something more like "loving relationhsip
with nobody" or "loving relationship with my dog".

So you start to do things that, you hope, will influence the state of the
"girlfriend" perceptual variable, moving it from "nobody" to "Michelle
Pfieffer". Since you (like most guys) have no clue about how to influence
this variable, you consciously try different things that you've heard of (go
to singles bars, become a rich movie mogul, drive a Ferrari, learn PCT, etc);
that is, you select references for other perceptions in the hope that
by bringing those perceptions to their reference state you will be able to
move the "girlfriend" perception towards its reference.

So you are always getting "feedback" about the state of the girlfriend
variable; you know if you have a relationship with a girl and you know
whether she looks more like Michelle Pfieffer or Mother Teresa. You also
"know" whether the "girlfriend" variable is moving closer to or getting
farther from its reference; this "knowledge" is embodied in the error signal.
The error is increasing if you are dating Mother Teresa and it's decreasing
if a beautiful girl just said "hi" to you.

If the controller doesn't have good control over the "girlfriend" variable --
that is, if there is no connection between error and action that reliably
brings the perception of "girlfriend" toward its reference level -- then the
error signal probably just drives the rate at which the controller randomly
changes from one "girlfriend getting" action to another. This is the E. coli
approach to getting a girlfriend. And it works. Although you might not end up
with "Michelle Pfieffer", you will end up with something close (in my case,
I came up with something that was not only close but even better; my wife.
And it was definitely a random walk getting there;-)

Best

Rick

[Stefan_Balke]

From Stefan Balke (960301.17.00 MEZ)

Rick Marken (960226.1230)

When you start controlling for a particular (reference) state of the
girlfriend perception (let's call that reference state "Michelle Pfieffer")
you are probably perceiving something quite different than the reference
state of the perception - - something more like "loving relationhsip
with nobody" or "loving relationship with my dog".
So you start to do things that, you hope, will influence the state of the
"girlfriend" perceptual variable, moving it from "nobody" to "Michelle
Pfieffer". Since you (like most guys) have no clue about how to influence
this variable, you consciously try different things that you've heard of (go
to singles bars, become a rich movie mogul, drive a Ferrari, learn PCT, etc);
that is, you select references for other perceptions in the hope that
by bringing those perceptions to their reference state you will be able to
move the "girlfriend" perception towards its reference.

I really like Michelle Pfeiffer and agree that I would do somethings like
that, but you say:
You consciously try different thinks... you select references...

My question is: How do I select them? Is there a model that describes the
process of selection?

If the controller doesn't have good control over the "girlfriend" variable --
that is, if there is no connection between error and action that reliably
brings the perception of "girlfriend" toward its reference level -- then the
error signal probably just drives the rate at which the controller randomly
changes from one "girlfriend getting" action to another.

Ok, this seems to be the answer to my question, but I expected something
others than just a random process. Is the consciousness nothing more like a
random generator?

This is the E. coli approach to getting a girlfriend. And it works.
Although you might not end up
with "Michelle Pfieffer", you will end up with something close (in my case,
I came up with something that was not only close but even better; my wife.

Congratulations, the same is true in my case

I don´t know much about the E. coli approach, except the demo, but there is
a german children game which came up to my mind while watching the demo.
Maybe it is also known in America, in German it is called Topfschlagen (pot
knocking) and it is usually played at children birthdays. At first, one
child out of the birthdays guests is selected to do the pot knocking. Then
that childs eyes are darkened by a shawl, so that the child now is unable to
see anything. Afterwards it is turned around itself to confuse it about the
own position in the room or garden and it is given a long wooden spoon. In
the meantime the pot had been located anywhere in the room or garden. The
pot is turned around and covers a little gift for the pot knocker. Now the
child tries to find the target (the gift under the pot) by a kind of guided
random walk. Because the child can´t see anything, it crawls upon the knees
and moves the spoon in all horizontal directions to find the pot. The other
children cry and scream hot, if the pot knocker comes nearer to the pot, or
cold, if the distance increases. Because the feedback always has a certain
amount of time delay (the arm and spoon movements are mostly rather quick)
and because there always a number of different children reporting not always
the same direction, the feedback is not really reliable. Additionaly it
seems to me that some children always create own ideas about where to search
best, or are to confused to follow the feedback, at all it lasts a while
until the spoon knocks the pot and the gift is found. The moves of the
searching child look very similar to the moves of the E. coli in the demo.

Regards, Stefan