[From Bill Powers (920923.0900)]
What a fine lot of posts this morning!
Penni Sibun (920922.1400) --
I'm not as far from your point of view as I may sometimes seem. I
think that direct private experience is the final arbiter. Any
theoretical idea has to explain EXPERIENCE. This is true even in
physics. The more points of contact there are between theories and
experiences, the more we can trust the theories to carry us over the
gaps in experience. So yes, intelligence, in the final analysis, is
what a human being experiences as intelligence. You can define it
objectively only through a theory that others, you hope, may accept.
That means, of course, that there are as many brands of intelligence
as there are observers. This can lead to pointless arguments, or it
can lead to an effort at communication, in which each observer tries
to say what she is looking at and how she is interpreting what is
seen, always with the humble awareness that one's point of view is
private and basically incommunicable. I suppose that's why we talk so
much.
Your criterion for determining sex, you say, is anything that can be
seen without a microscope. That's not quite saying "I know it when I
see it," because the criterion implies enumerable characteristics, no
one of which by itself is necessarily sufficient. And it doesn't
really handle behavioral aspects of sex, on which considerable
segments of the population don't agree. What is the sex of a nominally
male homosexual who is playing the cultural role of a female with
respect to his-her partner? Pat Buchanan, I'm sure, wouldn't bend his
definition of "male" to accomodate such a situation, but clearly the
male partner does, and those who have homosexuals for friends accept
the other person as the other person wishes to be defined. I don't
think that "I know it when I see it" is really sufficient to cover the
territory.
You say "organisms just don't go around doubting things most of the
time." I'll go further: Organisms don't generally realize that there's
anything to doubt. The curse of the theoretician is the discovery that
all is not as it seems, which puts an end to living a simple life in
which appearances are accepted at face value. It isn't that
appearances themselves are doubted, but once you begin to wonder why
things appear as they do, you discover the gaping holes in our
understanding and begin to see that the whole structure is a lot more
rickety than you would like. For Pat Buchanan, there are no holes; for
him life is simple. Why don't I envy him?
how does finding out whether the entity is a human or a program
have anything to do with intelligence as you suggest defining it
immediately above??
I was being a little sarcastic, suggesting that people generally
define intelligence to their own advantage and as a way of putting
down others, illustrating by showing how I would do that.
RE: the prickly correspondence between you and Rick. Why don't you
both lower the barriers and write down exactly the system concept you
have formed of the other person? Seeing it in print in a public place
might make it look a little different.
RE: turing contest
half the judges mistook the winning prog for a human.
Since there were only 2 humans, this meant that half the judges
mistook at least one human for a machine. To some people, although
possibly not you, the Turing test is a test of a program, to see
whether it exhibits an essential human characteristic labelled
"intelligence." The assumption is that if the entity at the other end
of the teletype DOES exhibit that characteristic, a human being will
know it by seeing it. The concept behind the test is greatly weakened
if it turns out that the human judge does NOT recognize this
characteristic in an interaction with an entity that the judge would
later agree is a certified human being. Certainly negative judgements
-- that a program has failed to capture this characteristic -- are
invalidated. And by implication, positive judgements, too, are cast
into doubt.
If the judges were AI experts and the audience was not, the superior
performance of the audience shows that the criteria the judges were
using were irrelevant to the characteristic being identified. So I
stand by my statement that AI types don't know what they're looking
for: they DON'T know it when they see it, even in the private terms
they would use when the nature of the entity is revealed to them.
ยทยทยท
------------------------------------------------------------------
Isaac Kurtzer (920922) --
Give your machine a kick for me. I think your analysis of Sparta is a
brilliant idea. Now that Tom's gone, I visualize you with a chair in
one hand and a whip in the other, surrounded by slavering toadies of
the psychological establishment snarling and swiping at you from their
pedestals while you force them one at a time through their hoops. When
you take your bow, watch your back.
-------------------------------------------------------------------
Pat Williams (920922.0900) --
Hi Bill. I'm sick of all this talk of me (Pat) and the feeding of
"healthy food" to my kids, so I'd like to try to clear up the
confusion about this particular instance of what Greg has been
calling purposive influence, myself.
I wondered how long you could stand it.
Yes, I am taking care NOT to disturb what the kids are controlling
for! That's the whole point. And I am also achieving my goal: they
are eating (not just being served) "healthy" food. Everyone is
happy! It is purposive because I'm doing it on purpose; and it is
influence because the kids are acting differently (eating "healthy"
food) than they would be if I served them just any old tasty food
or let them scrounge for themselves (then they would be eating
tasty food, most of which would not meet my definition of
"healthy").
We agree that you're controlling your own perceptions, on purpose. We
disagree that your kids are acting differently because of your
controlling for nutritional value of the food. What they're getting is
different because of that, but what they're controlling for isn't.
Their actions would remain the same no matter what you served them, if
they ate it. What we're trying to sort out here is not whether you're
doing the right thing by your kids but who is influencing and
controlling what.
You would be frustrated if your kids didn't eat what you served, but
the eating is still up to them, not you, even when they eat it. You
can't operate their arms and mouths and swallowing muscles. You may be
gratified at seeing that good food disappear irrevocably into their
mouths, but you can't influence that process or control it: matters
have passed outside the sphere of your influence at that point. Your
influence has reached its limit when you put the food down in front of
them.
What I'm talking about is much simpler than the arguments we're going
through. It has nothing to do with your feelings about what the kids
are eating, your reasons for giving them the food you give them, or
what the kids would do if you behaved differently. It has to do with
who is in charge of which perceptions and which means of controlling
them. The kids are controlling their own perceptions. You are
controlling yours. As a result, there are certain interactions between
you and the kids -- disturbances, or if you like, influences. These
interactions come about because when you do what's necessary to
control your own perceptions, your actions disturb the kids' worlds,
and they adjust their actions to prevent those disturbances from
having any significant effect on the variables THEY are controlling,
if they physically can. Just try helping them put the food in their
mouths: reach out, grab a wrist, and push the fork toward the mouth.
They will resist. Even helping is a disturbance if it's applied to
something that's already under control.
I get the impression, possibly unjustified, that you and Greg are
looking for something more than that. You don't want to see this
interaction as just "disturbances," or as just neutral "influences,"
but as a way in which one organism can have an intentional effect on
the interior life of another organism. In other words, you want to
think you're doing some good for your kids, not just going through the
motions and having the kids' control systems counteract everything you
do. This has got to be a sensitive issue; you go to great trouble to
do the right thing for your kids, even to the time-consuming labor of
schooling them at home. And here comes Bill Powers saying that all
you're doing is disturbing them, and that all they're doing is making
sure your disturbances have no effect on them. Makes it all look
pretty futile, doesn't it?
But I don't think we can come up with a correct understanding of human
relationships unless we're willing to face the (seemingly) worst
possibilities and put them aside. Perhaps it will turn out that your
efforts with your kids aren't futile. That would be nice. But I think
it would be better to know the true relationships than to prove that
you've been right all along. It's very difficult to come up with a
valid explanation or analysis if there are possible outcomes that are
ruled out from the start, and others that MUST be supported. This is
why I have tried to point out to Greg that he has a GOAL in this
argument; that he's trying to make it come out a particular way. It's
trying to make an argument come out a particular way that leads to
switching definitions of words, generalizing upward by one set of
rules and particularizing downward by a different set, and all the
other slippery ways that human beings have of directing their own
reasoning toward a conclusion that they want to be true. If you want
to know the truth, you can't care how the argument comes out. You just
have to follow it through and see where it leads, trying not to make
any mistakes.
I think there are ways in which we can intentionally influence the
interior lives of children. I don't yet have a clear picture of what
they are, and when the answer gets closer it may turn out to be the
opposite of what I think. But at the moment I think that because
children demand instruction, we can give it to them. We can't do the
reorganizing for them; if they don't do it, it won't get done. We
can't control how this reorganizing will come out. But we can make
sure that the environment contains everything we know about that might
be useful to them, and that what we tell them about is as near to true
as we can approach. We can invite their attention to things,
relationships, problems that they might not come across by themselves.
We can look for difficulties they are having, and demonstrate how
another person would deal with them. We can make sure that when
they're ready to perceive in a new way, there are things to perceive
at the new level that will be generally useful to perceive -- although
of course we can't say just how they will end up perceiving the
situations we perceive.
Chuck Tucker has the right view, in my opinion. We don't give
perceptions or reference levels to others; they TAKE them. They suck
them in, when they're children. Whatever they've taken in, they work
over, rearrange, test, try out, think about, and convert into
something they, not the external world, find useful.
-------------------------------------------------------------------
Chuck Tucker (920922) --
I agree with you so much that I have nothing to say to you. Hello.
Hope you're feeling fine.
-------------------------------------------------------------------
Greg Williams (920923) --
All that can be influenced is the action used to correct any
induced error, and that can be influenced only in the sense of
making more or less of it necessary.
Yes. Well, actually, "more or less" in a not strictly real-number-
line way, but possibly very complexly "qualitative," being
multidimensional. Even in plain old rubber-banding, the actions
needed to compensate for the applied disturbances are not simply
"more or less" in one dimension, but simultaneously in TWO
dimensions.
"More or less" in each dimension being controlled and disturbed,
you're right. In the rubber-band experiment on a tabletop, you can
disturb independently in two dimensions. The other person controls
independently in two dimensions (possibly two different dimensions,
but two). The "more or less" applies in each dimension.
The nature of the action is already determined by the organization
of the control system, which takes into account the environmental
link through which action affects perception.
Take another look at the generic PCT diagram. The action is
determined conjointly by BOTH the reference signal AND the
disturbance, via subtraction, one from the other, at the
comparator.
Take another look at the statement. The NATURE of the action means
what KIND of action has to be produced, not how much of it. The KIND
of action is not determined either by the reference signal or by the
disturbance. It's determined by the physical links that exist between
the system's output and the controlled variable, by the perceptual
function defining that controlled variable, and by the effector design
of the control system. The reference signal and the disturbance
determine, conjointly (if that means anything) only HOW MUCH action
will occur, and IN WHAT DIRECTION. They do not determine WHICH KIND of
action will occur.
You can claim credit for those increases and decreases, but not
for the nature of the action. That was established before you
arrived on the scene.
"You" cannot claim FULL credit. Only SHARED credit. SHARED with the
organism. The nature of possible actions, given possible
disturbances, WAS established by the organism, but YOU select WHICH
of the possible disturbances ACTUALLY is applied.
The organism determines WHICH controlled variables it will be
controlling at a given time. You can then pick any of them to disturb.
When you do, the organism will vary its action on that controlled
variable, cancelling your disturbance. It is, of course, already
varying its actions on ALL the controlled variables, or is ready to do
so, at all times, if they're under active control. You're not
determining which variables will in fact be under control or which
actions they will use. That's been decided already.
Our image of "rubber-banding" is unfortunate in one respect, because
this demonstration has deliberately been made very simple, to
illustrate principles. A more realistic example of rubber-banding
would give the control system one rubber band attached to the knot,
and twenty different people twenty rubber bands attached to the same
knot. The control system won't have any difficulty in controlling the
knot (unless the combined disturbance results in breaking of the
control system's rubber band) because only the vector sum of
disturbances matters. Control might actually be easier because
independent random disturbances will sum to a net disturbance having
much less variability than any one of them has.
But any one person acting as a disturbance, trying to influence the
control system's hand position, is going to have great difficulties
because of all the other random disturbances that are present. While
control still remains possible, it's no longer possible for the
disturber to estimate the best direction to move his/her own hand to
achieve a correction of the other's hand position, because there is no
longer any best direction. And it becomes difficult for the putative
disturber to know what disturbance is actually being applied;
perceiving one's own rubber-band tension is no longer indicative of
the net disturbance on the other's controlled variable. The only way
to make sure of applying a known disturbance is to isolate the control
system from all those other influences.
There's no difference in the action used to counteract an
accidental influence or an intentional influence.
Correct. But that is beside the point. Purposive influence matters
to the influencer ALWAYS.
Agree. Control system control their own perceptions. Are you making
some new point here? Why the "but?"
I think I now am beginning to understand how your tenacious
clinging to the organism's point-of-view, which serves you so well
scientifically ... seduces you into an ideological position at
odds with PCT-science.
The organism role IS more important than the environment's role to
the psychologist who wants to see order (invariable goals) behind
the chaos (variable actions). But it is a mistake -- because it
contradicts PCT-science -- to carry that importance over to
ideology.
Are you saying that a psychologist has a point of view that is not
that of an organism? Good trick. I'd like to learn how anyone can have
a point of view different from that of an organism. Go ahead -- I'm
easily seduced.
------------------------------------------------------------------
As I understand your goal in this argument, it is to prove that
the environment has influences on the organization of the
organism's control systems, shaping the organism in some causal
way. This goal includes proving an effect of both non-living and
living influences on the organism, independent of its heredity.
Yes, that's my aim in the second of our two arguments. The first
argument involves purposive influence only, not environmental
influence in general. Of course, they have been intertwined
somewhat.
Well, that means we should be able to save a lot of time. If proving
this conclusion has been your goal from the start, you must have had
other reasons for believing it is the correct goal, reasons that go
beyond, predate, or supersede PCT. This would, of course, explain why
you don't develop an irrefutable argument based only on PCT
principles: there are other principles involved. It would also explain
why, when you misinterpret a term like "nature" (above), you always do
so in the direction that leads to your goal, and never accidentally
the other way. When you are convinced that there is only one valid
outcome of an argument, which you already know, the details of the
argument cease to be important. If one argument bombs out, there's
bound to be another that will get you there.
In my lexicon, knowing the answer and coming up with whatever argument
it takes to get there is called "ideology."
--------------------------------------------------------------------
Whatever the people do, they will do it with a particular
organization of control systems, and exercising those control
systems or bumping them up against each other is not going to
change anything.
Look around. Peoples' lives are changed in both trivial and
profound ways (as judged by the influencees themselves) everyday,
all over the globe, by acts of purposive influence. Ask THEM
whether "anything" has changed.
If they say anything has changed, they are ignorant of history. What
you speak of is not change; it's just coping. Political intrigue is
the same today as it was in Rome or Egypt. People try to push other
people around just as they have always done. The interactions and the
influences fluctuate and shift around in complex patterns that achieve
no more today than they have ever achieved, with the exception that
now there is enough power available to achieve extinction. Without a
correct understanding of human organization, human beings simply cope
with whatever immediate problem presents itself, only dimly aware that
there are others doing the same thing, and under the false impression
that they're accomplishing something of lasting worth instead of
causing most of their own problems. It may well be that there is a
slow drift toward understanding. But it's too slow to suit me.
What you present as "purposive influence" is simply control of others.
You may say it isn't, but you still insist that purposive influence is
intended to have some effect other than just disturbing what others
are controlling for. Look around. See what it has accomplished.
Perhaps you would get a better view from Chicago, or Los Angeles, or
Albuquerque. What I see is not just a natural progression of
interactions. I see the influence, if you will, of a widely-held
misconception, a disastrously wrong conception, of what a human being
is and how a human being works. That's what has to be changed -- not
what we "do."
------------------------------------------------------------------
Best to all,
Bill P.