SV:behavior

[R_Marken]

[From Rick Marken (990319.0830)]

Bjoern Simonsen (990318.24:00 EU-time)--

For mee it is important to mark that behavior is the modification
of an object and no more.

Why is that important to you? If you are already committed to
some ideas about how behavior works I think it's going to be
very hard for you to learn PCT.

Assuming for the moment that you do want to learn PCT, I will
just note that what you modify about an object when you control
it is some _variable aspect_ of that object. For example, my hand
is an object that I can "modify". But it is actually _variable
aspects_ of my hand that I modify: I can move my hand (vary its
position), make it clenched or open (vary its shape), keep it
steady or waving (vary its velocity), etc. Position, shape, and
velocity are _variables_. What we control (maintain in predetermined
states) are aspects of the world that can _vary_; we control
_variables_.

Does it bother you if I say we control _variables_? It may
be possible that we have a language problem here. Perhaps
what you mean by "object" is what we mean by "variable". There
is no question that, in English, the only appropriate way
to describe control is as the process of bringing _variables_
to predetermined states and protecting them from disturbance.
Do you, perhaps, mean "variable" when you say "object"?

Me:

How about "behavior is the control of perceptual variables"?

You:

No, I prefere
_Behavior is any active and purposeful modification of an object,
detectable externally_

But you can detect "externally" (as an observer) the perceptual
variables that a person is controlling. Control of perception
is a perfectly objective phenomenon. Indeed, it's more objective
than just looking at behavior and "detecting" what the person is
doing; what you "detect" may an irrelevant side effect of what
a person is actually doing (intending). For example, you may
notice that a person is making an obscene gesture with his
hand; this is a "modification of an object" (the person's
hand) that you can detect. But this obscene gesture may be a side
effect of the person controlling some other variable (such as
the degree of itch produced by a piece of dust in the eye).
The obscene gesture is an irrelevant side effect of scratching
the eye (controlling itchiness).

Re: The Test for the Controlled Variable

I understand, but are you sure that the computer controls
the accelaration of the squares?

The computer doesn't control anything this demo.

I think the computer controles some of the modifications in
the group of sqares.

It doesn't. The person controls (if they want to) the position
of _one_ square at a time. The person's actions _influence_
all three squares. That's why it's hard to tell (by looking
at the movement of the squares) which square is being moved
_on purpose_. The computer does The Test for the Controlled
Variable on _all three_ squares continuously and simultaneoulsy.
It takes a few seconds for the computer to figure out which of
the three squares (if any) is being controlled at any time.

Rick, I found the program for some of your demoes. But I
didnt understand the word ASAP.

The Java source for the "Test for the Controlled Variable"
demo is now up. ASAP is an abbreviation for As Soon As Possible.
I meant to get the program up on the net ASAP and I did:-)

Before we continue this conversation, Bjoern, perhaps you could
let me know whether you actually want to learn PCT or not. If
you are already committed to some idea about how human
behavior works (if, for example, you are committed to the
idea that behavior is externally observable modifications of
objects) then it's really a waste of your time and mine to
have this conversation.

Best

Rick

···

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

[T_Harms1]

[from Tracy Harms (990319.0940 Pacific)]

Rick Marken (990319.0830)to Bjoern Simonsen:

... what you modify about an object when you control
it is some _variable aspect_ of that object. For example, my hand
is an object that I can "modify". But it is actually _variable
aspects_ of my hand that I modify:

The short form of my previous message is: Good job, Rick. I think this
manner of explanation shows refinement in your ability to communicate
Perceptual Control Theory. Let me add to that my opinion that I've long
considered you among the best propagators of PCT.

Now we'll just have to examine how well the use of this phrase actually
serves in practice.

Tracy Harms

[R_Marken]

[From Rick Marken (990327.0940)]

Me:

The study of perceptual variables is economical (if not easy)
because it requires only one piece of lab equipment: the mind.
It's the kind of research everyone can do, even those who don't
want to do research.

Bjoern Simonsen (990327;09:50 EU-time) --

I dont agree.

I need an instrument to detect that your beige computer changes
color from hour to hour ( if there is a window in your computer
room) I also need an instrument to detect the modifications of
my blood pressure over time.

Instruments just provide perceptions. What you perceive when you
look at a volt meter, say, is _not_ a flow of electrons; you
perceive the _position_ of a _pointer_ relative to _numbers_.

An instument can't tell you what I am perceiveing; only I can
experience _my perceptions_ of the color of the computer. For
that matter, only I can experience _my perceptions_ of the
readings of the instrument that purportely detects the changes
in the color of my computer that I don't detect.

Rick. Have you ever thougt about a world consisting of objects
where there are no modifications (at all).

Not much. The fact that perceptions change (vary), often in ways
that move them from our references specifications for those
perceptions, is the reason _why_ living systems _are_ control
systems. Living systems require that certain perceptual aspects
of their internal and external environments remain in specific
(reference) states.

Me:

Have I modified an object when I change the relationship between
cup and fork from "next to" to "behind"?

Bjoern:

Yes, if the object consists of the system cup and forc.

Ok. Have it your way:-) If you are able to talk about variations
in aspects of these objects -- perceptual variables -- you still
will be able to understand the notion of "control of perceptual
variables", which is the central concept in PCT.

Me:

People can't really "give" feedback. But I don't want to even
start to get into _that_.

Bjoern:

May be an other time?

Oh, all right. "Feedback" is a functional relationship between a
variable and itself. It is not a "thing"; it is the name for a
particular type of causal relationship between variables,
specifically one that looks like this:

x-->y
^ |

___|

where --> means "causes". In this little diagram, variations in x
cause (feedback on) themselves via y. The causal relationship
between x and itself is called "feedback".

Bill wrote a definition in B:CP:
NEGATIVE FEEDBACK: A feedback situationin which a disturbance
acting on any variable in the feedback loop gives raise to an
effect at the point of disturbance which opposes the effect
of the disturbance.

Negative feedback is a particular type of feedback relationship,
one in which the feedback relationship between a variable and
itself causes that variable to go to zero. In a living system
negative feedback looks like this:
             r
             >
             v
         p-->x-->y
         ^ |
         > >
    d--> q<-------

The loop is negative because the variable x (the error signal)
causes itself to go to zero. d is a disturbance variable; note
that it is "outside" of the feedback loop; it is not in the loop
as y, q and p are. What Bill is saying above is that a negative
feedback loop acts (in this case by varying y) to to counter
the effect of d on a variable in the loop (q in this case -- the
controlled variable).

Variable r is also outside of the loop; it's effect on the behavior
of the loop is somewhat different that d's. A change in r (the
reference signal) leads to a change in p so that x is kept a zero.
The fact that a negative feedback loop forces p to track r while
preventing d from having any effect on q (and, thus, p) is the
reason why we say that the behavior of a negative feedback loop
is the "control of perception" (control of p).

Is there a feedback loop in the loop:

I would say that any causal loop is a feedback loop unless
there is zero loop gain (eee Powers' LCS I; the paper starting
on p. 129).

Am I wrong when I think about your comments as disturbances in
a feedback loop?

No. You are correct. My comments are probably a disturbance to
some variable you are controlling. My comments will almost
certianly have no effect on your perceptions;-)

And I am interested in PCT because I hope to find a model/theory
for purposeful and teleological behavior (Wieners way of using
the word teleological)

You've found it!

Best

Rick

···

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

[Marc_Abrams3]

from [ Marc Abrams (990327.1256) ]

Rick, Bjoern, Hope you don't mind my jumping into the thread, but I might be
of some help. ( Maybe ).

[From Rick Marken (990327.0940)]

Me:

The study of perceptual variables is economical (if not easy)
because it requires only one piece of lab equipment: the mind.
It's the kind of research everyone can do, even those who don't
want to do research.

Bjoern Simonsen (990327;09:50 EU-time) --

I dont agree.

I need an instrument to detect that your beige computer changes
color from hour to hour ( if there is a window in your computer
room) I also need an instrument to detect the modifications of
my blood pressure over time.

I think this might be the first hurdle. Bjoern, what is your definition of a
perception? Am I right in thinking you believe that there is an objective
world out there that we "perceive" (i.e. we "interpret", cognitively what we
see ). So each of us could "perceive" different "aspects" of the same
objective objects?, and "interpret" the world and it's objects differently?

I believe everything else from Bjoern follows from this basic view and
understanding

An example:
Rick:

Instruments just provide perceptions. What you perceive when you
look at a volt meter, say, is _not_ a flow of electrons; you
perceive the _position_ of a _pointer_ relative to _numbers_.

An instument can't tell you what I am perceiveing; only I can
experience _my perceptions_ of the color of the computer. For
that matter, only I can experience _my perceptions_ of the
readings of the instrument that purportely detects the changes
in the color of my computer that I don't detect.

Bjoern:

Rick. Have you ever thougt about a world consisting of objects
where there are no modifications (at all).

Well Bjoern, is my assumption accurate?

Marc