NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/bit/listserv/csgl/1812

Text File | 1992-12-12 | 11.5 KB | 246 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!paladin.american.edu!auvm!VAXF.COLORADO.EDU!POWERS_W
X-Envelope-to: [e-mail]
X-VMS-To: @CSG
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.csg-l
Date:         Sat, 12 Dec 1992 04:55:22 -0700
Sender:       "Control Systems Group Network (CSGnet)" <[e-mail]>
From:         "William T. Powers" <[e-mail]>
Subject:      Many subjects
Lines: 231

[From Bill Powers (921212.0300)]

Might as well get the mail and write about it. What else can you
do when you're awakened at such an hour to be told you have a new
grandson (my daughter Barbara's second)?
---------------------------------------------------------------
Len Lansky (921211.1143) --

You raise an excellent set of questions. At a more propitious
time I will start putting together an outline for a one-quarter
(13-week) course in PCT. Unless someone else beats me to it. I
envision a group project here, in which people rewrite the
outline and re-rewrite it until we have something that looks
good.
--------------------------------------------------------------
Rick Marken(921211) and Ed Ford (921210) --

RE: Science and religion

It will not be possible for science and religion to get together
until both realize that neither is Revealed Truth, and that both
are human ideas. Of course that is precisely what both sides have
been rejecting since the start of science. One side points the
finger at Nature, the other at God. Neither side, apparently,
notices whose finger is doing the pointing.
----------------------------------------------------------------
Gary Cziko (921210.1622) --

The Little Man starts out (in the mapped mode) with a "bilateral
lesion" that makes it consistently point short of (and to one
side of) distant targets. As it builds up the map it gets closer
and closer until it points exactly to the target on the first
try. I suppose you could put a "lesion" in the map that
selectively eliminates the correction in any dimension --
horizontal, vertical, or depth.
----------------------------------------------------------------
Bruce Nevin (921210.1206) --

Come to think of it, language IS like grooming. Sometimes you
pick up something nice and salty, but sometimes it wiggles and
quirms until you crunch down on it.
----------------------------------------------------------------
Oded Maler (921212) --
RE: why is a theory of behavior necessary?

Why is it necessary to have a theory of gravitation, or
electronics, or numbers, or anything else? I suppose "necessary"
is the wrong word here -- after all, other species have got along
fine without any such theories. One can certainly jump and fall
down without a theory of gravity, and behave without a theory of
behavior, or even with a wrong theory of behavior. But theories
in general seem to make sense of things in a way we can't achieve
by any other means.

Or were you specifically asking why we need a theory of behavior,
in addition to all the other
 theories?---------------------------------------------------------------
Penni Sibun (921211) --

RE: Not "believing in" purpose.

I was about to respond in a way almost identical to Rick's, by
saying that purpose is the essence of control. But you then went
on:

>i think this ``virtual synonymy'' is a rhetorical move of pct's
>that irritates people w/ a scientific bent (like me).

If you think that the PCT interpretation of the meaning of
purpose, or intention, or desiring, or wishing is a "rhetorical
move" then you simply haven't caught on to the basic architecture
of control.

>i find yr generalization to everything else, including the
>kinds of things implied by ``purpose,'' completely unwarranted.
>so i wish you'd stick to making claims you can justify....

What kinds of things do you think are implied by "purpose?" The
PCT claim is that purpose is a perfectly real phenomenon,
explainable by a proper model of control behavior. Purposive
behavior is behavior that results in bringing some aspect of the
environment or oneself or the relationship between them (as
perceived) to a predetermined state, and maintaining it there for
some time despite disturbances that tend to change it. This is
the phenomenon that science mistakenly rejected in the early
decades of this century, claiming that it required the future to
affect the present and a number of other foolish things. The
problem really was that science, at that time, lacked any
explanation for purposive behavior, and at the same time assumed
that what science could not explain (in, for instance, 1920) was
mysticism, superstition, or illusion. That opinion prevails
today, especially among scientists who still don't understand how
control theory explains purpose. There are always scientists who
think that what they can't explain in terms of the science of
their time (the part they know about) can't be explained at all.

Since you're the main person around here with a scientific bent,
how about giving us the really scientific explanation of how a
person carries out a purpose?
-------------------------------------------------------------
Wayne Hershberger (921211) --

>What is wrong is supposing that the conceptual EVs are not part
>of the model of the perceptual process!  The light-ray tracings
>in Bill's little stick man simulation comprise an integral part
>of what works!

There are others who would argue on your side -- some
cyberneticists, in fact, claim that not only are the light-ray
tracings part of the model, but that the outputs that alter the
light-ray tracings are part of it, too. Basically they argue
that any separation of the organism from its environment is a
conceptual mistake. It's all just one big system, so there's no
point in trying to take it apart into components in order to
understand it. You can only understand, like, the WHOLE THING.

I for one have never found that contemplating the WHOLE THING
leads to anything but bafflement. I think that many people
confuse this sense of bafflement with a sense of knowing.

The light-rays are not only part of the perceptual processes,
they are part of the output processes. We could just as well say
that the actions of the Little Man consist of altering retinal
images, or to go further that they consist of alterations in
neural signals in the brain. In other words, it's all output,
not perception. When you trace out the signal paths, you find
that what we mistakenly call perceptual signals are really just
part of the whole process of output.

When you go far enough with this, you must finally decide that
the outputs really consist of outputs, because when you start
with the motor output forces and keep adding all the things
directly related to and dependent on them, you end up back with
the output forces after one trip around the loop. Making all the
substitutions to eliminate intermediate terms, you end up with
output = f(output). Or perception = f(perception), or error =
f(error). It all depends on where you start.

In our modeling efforts, we have found it definitely useful to
distinguish the organism from its environment. When we do so, we
find that the organism can know only what its senses tell it. And
we find that what its senses tell it about the environment
depends critically on how the sensory processes are organized.
Two organisms in the same environment experience two different
environments. Yet the mere fact that we can say "the same
environment" means that we suppose that something does exist
there independently of the perceiver. We can't prove that, but we
can build models on the assumption that it's true.

>... perceptual EVs and the conceptual EVs are lawfully related
>to each other (I gather we agree fully on this point).

Yes, but I doubt that you will agree with my agreement. I
consider conceptual EVs to be higher-level perceptions of lower-
level perceptual EVs, and that the lower-level perceptual EV's
are in turn functions of variables in the Boss Reality. The
higher-level "conceptual" EVs consist of such things as
relationships, categories, sequences, logical and syntactical
propositions, principles, and system concepts. Each successive
level consists of perceptions that are functions of perceptions
of lower level. The "lawful relationship" of which you speak is
simply the relationship in which higher-level perceptions depend
on lower ones according to the form of the higher-level
perceptual function. If you look at the content of any
"conception," you will find elements that correspond to the
levels I propose -- and, I claim, nothing else. If you do find
something else there, please let me know.
I disagree, therefore, with the distinction between perception
and conception as being too crude. I use the term "perception" to
mean any afferent process at any level -- that is, anything we
can experience. In place of the simple dichotomy perception-
conception I substitute specific levels of apprehension of the
world, and make the claim that these levels are hierarchically
dependent. The general achitecture of the brain, and the effect
of lesions and ablations in various parts of the brain, supports
the view that the "conceptual" levels have access to the world
only through lower levels of input processing. The highest levels
have no direct access to the world -- there is no uninterrupted
input path from the sensory organs.

>It is this lawful relationship that allows us to use our
>perceptual EVs to test our conceptual EVs in scientific
>experiments.  Science does not address the question of whether
>our conceptual EVs (making sense of the immanent order)
>correspond to those of some divine conceiver (God or Boss).

You are fighting a straw man here. To hypothesize that there is a
Boss Reality is not the same as saying that there is a Divine
Conceiver. I'm surprised that you keep bringing this up -- surely
you don't think that this is my explanation of the Boss Reality!
You seem to be ruling out the possibility that there could simply
be a universe that is independent of our experiences but on which
our experiences depend. Or you seem to be saying that the only
way such a universe could exist would be in the mind of a
supernatural being.

This truly puzzles me. I don't see how you arrive at just those
possibilities. The only possible clue I can get is in this:

>That relationship [between concept and percept] is the basis of
>empiricism.

and this:

>Not only does the immanent order have many more degrees of
>freedom than we can perceive, it may have more degrees of
>freedom than we can conceive.

It seems clear to me that what you are calling the "immanent
order" is what I call the "Boss Reality." Models are a way of
conjecturing about what might exist in the immanent order that we
have not yet perceived. We imagine, in other words. We provide
for ourselves lower-level perceptions synthesized from memory,
and view them from higher levels as if they exist in real time.
We categorize and characterize them, we reason about them, we
form principles on their basis, we fit them into a systematic
concept of the world. And we test them: we create outputs
calculated to produce certain perceptions after passage through
that external world as we model it, and check to see if the
result is what we expected.
Of course we must always cast these conjectures in the form of
human perceptions. We have no other way of experiencing the boss
reality.

Perhaps all of this comes down to the same concept that you call
"empiricism." But I think you may have more faith in empiricism
than I do, as a source of true knowledge.
----------------------------------------------------------------
Best to all,

Bill P.