NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/bit/listserv/csgl/1431

Text File | 1992-11-09 | 7.4 KB | 152 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!stanford.edu!bcm!convex!darwin.sura.net!paladin.american.edu!auvm!AERO.ORG!MARKEN
X-Delivery-Notice:  SMTP MAIL FROM does not correspond to sender.
Posted-Date: Tue, 10 Nov 92 10:39:55 PST
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.csg-l
Date:         Tue, 10 Nov 1992 10:39:55 PST
Sender:       "Control Systems Group Network (CSGnet)" <[e-mail]>
From:         [e-mail]
Subject:      means-ends;language
Lines: 139

[From Rick Marken (921110.0900)]

Bill Powers (921109.1430) --

Excellent. Beautiful.

I, of course, did not mean to imply that I thought that the
ends justify the means. I was just making a point about control;
any controlled result is influenced by the controller and independent
disturbances. So, in order to produce a consistent result (end) the
controller must vary his/her influences (means) so as to counteract the
influences of the disturbance. We all know that controllers
can do this because they are organized as negtive feedback systems
which keep a perceptual representation of the controlled result
matching a secularly adjustable reference input (thanks for the
GREAT tutorial on PCT modelling, Bill; I'll send a BASIC and pascal
version ASAP).

I suppose the article of "faith" (as Martin puts it) of PCT is
that ALL controlled results are perceptions that are controlled
in this way. This includes complex results that (in Bill's model)
are called "programs", "categories", "principles" and "system concepts".
It is difficult to deal with these variables quantitatively, so
there is plenty of room for argument about what is being controlled
(in language, for example) and whether it is controlled (since it
is difficult to perform "the test" at these levels).

The fact that higher level variables are controlled has to rely,
for now, on somewhat more subjective evidence. For example, we
know we control something that might be called "sentence structure"
(at least when we write) because we can tell when a sentence seems
well constructed and when it doesn't; and we (some of us) will
rewrite sentences to make them sound better -- ie. we control something
what could be called "sentence structure".

So it is a jot more than faith that leads us to assume that things
like programs and system concepts are controlled -- we can experience
errors at these levels and we can (sometimes) see people (or ourselves)
acting to correct these errors. In order to control these higher level
perceptual ends we MUST be able to adjust references for lower level
perceptual ends (which are means with respect to the higher level
ends). I think this must be considered a fundemental fact of control
system operation -- as much as the fact that control systems control
their perceptions.  This is why I think Ed Ford (921110) is wrong when
he says:

>It seems to me that by forsaking our own systems concepts, that which
>we've built over a life time of thinking and experiencing, for the sake
>of what we perceive as social demands or social approval, as More
>states, we lead our country as well as oursevles "by a short route to
>chaos."

"Forsaking our own system concepts" suggests that we are able to
vary our references for our system concept perceptions. For example,
instead of having a reference for "being Catholic" we forsake this
system concept by setting it to "being liberal Catholic". What else
could "forsaking a system concept" mean in the context of HPCT?
Variation in the reference for a system concept (in the current version
of HPCT) could happen in 3 ways: 1) as the result of random
drift (not a plausible option, really) 2) as the result of
reorganization due to intrinsic error and 3) as a result of the
disturbance resistence being carried out by higher order systems.
The first option suggests that references for system concepts change
randomly -- and I can't believe this would work in a control organization.
The second two options seem most reasonable -- but in both cases,
"forsaking of systems concepts" is DEMANDED in order to maintain
control; in the first case, the system concept is forsaken in order
to maintain the viability of the organism -- to eliminate intrinsic
error; in the second case, it is done to compensate for a disturbance
to higher level controlled variables. But in both cases the system
concept is a means for controlling another perceptual variable.

So a person must (and will) forsake system concepts in order to control
other variables (if system concepts references can be varied; but obviously
they can); this kind of flexibility is not a problem; it is
a requirement for control. In fact, when references which can be
varied are fixed at some level "no matter what" it is a symptom of
conflict (as you mentioned later in you post, Ed); and conflict is
the enemy of control. When a person controls for beliefs that
create conflicts (like the belief that you are a sinner if you
don't worship jaweh) then they are limiting their ability to control.
I think that people have been far too UNWILLING to forsake some of
their system concepts, principles and programs -- much to their
own unhappiness (from lack of control due to internal conflict)
and that of others (due to lack of control due to efforts to make
them have the correct beliefs -- interpersonal conflict).

The control of perception implies flexible means to control
perceptions -- unless, as Bill noted, you happen to be rich and
well armed. And beliefs -- blind acceptance of certain reference
levels for system concepts, principles, programs, etc -- are the
enemy of flexibility because they create conflict.

PCT slogan # 2:

Belief is the enemy of control.

penni sibun (921106) --

>i certainly am not interested in your judging the value of my model.

That's fine. I'll stick to PCT.

I will just say that, in the basic sketch of your model, the PCT
aspects are quite clear:

   > another way to characterize this might be 1) where are we in the
   >linguistic structure (ws) and 2) where are we in the structure of
   >``meaning(s)'' (?) we are talking about.

From a PCT perspective,  linguistic structure and the structure
meanings are controlled variables; so they are perceptions with
intended (reference) states. That is how your model relates to PCT. I
think it would be possible to look in your model and (to the extent that
it controls) see what corresponds to the perceptual representation of
these variables and their reference specifications. But it seems to
me that you are not interested in really trying to see the controlling
aspects of your model -- and what it does. So that's fine. You seem
to get a little angry at our suggestions (which, at least in Bill's
case have been made quite clearly) -- and I don't think there is
anything close yet to the kind of data we need to do PCT modelling
of language phenomena. I think language phenomena are VERY interesting
but I don't think we'll make much progress on it until someone who
really understands PCT starts doing the kind of studies and modelling
that needs to be done. I don't have the time and obviously you don't
have the inclination. So let's stop fighting. You're happy doing language
modeling your way -- go for it!

Best regards

Rick

     **************************************************************

Richard S. Marken                   USMail: 10459 Holman Ave
The Aerospace Corporation                   Los Angeles, CA 90024
E-mail: [e-mail]
[phone] (day)
[phone] (evening)