[From Bill Powers (920906.0700)]
Greg Williams (920905) --
Greg, you keep wanting to make me say that people can't influence each
other, when what I'm saying is that they can't control each other
(except with respect to output). The confusion is coming partly from
your using influence not just to mean "have some effect on" but in the
sense of "have a particular preselected effect on," which is not
influence, but control (that is, it can't be done by any means but
control).
I want to call attention to a different sentence in the paragraph
which >"seems fine" to you: "Another person's control structure
(reference >signals) cannot be so controlled regardless of the threat
(or perceived >threat) of or direct use of overwhelming physical force
by you."
I had missed the significance of the parenthesized (reference
signals), which as Rick Marken pointed out is not the same as a
control structure. A control structure is (to me) a physical
structure, not the signals being passed around in it. But that gets
ambiguous, so let's let that go, too.
I think you're missing the point of "overwhelming physical force." If
I resort to that, then it doesn't matter what you do or don't do. I
may try a threat first. If that doesn't work, I can try hurting you.
If that doesn't work, I can try seizing you and forcing your limbs to
move as I want them to move. As long as I have superior resources, you
WILL do what I want, even if I have to do it myself using you. As a
final resort, I will just kill you and go looking for someone easier
to control (why work up a sweat?). To a controller of people, people
are interchangeable. You're thinking of the use of force as being
constrained by some kind of civilized limitations. Unless one gives up
the goal of control, there are no such constraints. You do whatever it
takes, to quote a prominent humanitarian.
(1) Consider again my yelling-"fire"-in-a-crowded-theater scenario
which I set forth earlier this week. I yell "Fire!" and you (being
prudent -- that is, in order to satisfy a reference signal in you
having to do with not getting burned) move toward the exit. I am
controlling for my perception of seeing you move toward the reference
signal, so if you didn't move thusly, I would alter my actions so as
to >(I hope) see you move thusly. The situation as construed would
appear >to fit my definition of "manipulation" -- I am controlling
your >actions, as per the first sentence in the paragraph above which
you >OKed. But I claim that I am also altering your control structure:
undeniably, it is operating differently than it would have had I not
yelled "Fire!" We can mince words about whether your reference
signals >per se have been altered, but it is clear that the operations
of your >control structure are different than they would have been
otherwise.
It is undeniable that my control systems are operating differently in
the presence of a disturbance from the way they would be operating
without the disturbance. There is nothing different about their
organization, of course, but only in the signals that are moving
around in the system. The way you're arguing, that would mean that ALL
disturbances are "controlling my control structure."
I don't need to mince words about whether reference signals have been
changed. They have. I can't lift a finger without changing reference
signals. I can't resist gravity without resetting reference signals
when I change position. For ALL systems in the hierarchy, acting means
changing the reference signals of lower systems. There are no FIXED
reference signals in the whole (healthy) system, except perhaps at the
highest level -- and even those are subject to change, I think. I
hope.
When you yell "FIRE" you insert a disturbance into my system at some
level. You do this by altering an uncontrolled perception that is a
component of a higher-level controlled perception. To combat this
disturbance, I naturally have to shift reference signals at all lower
levels. That is how action works. I alter the controlled lower-level
perceptions (by means of altering reference signals) so that when they
are put together with the uncontrolled one you have changed, the net
result is no disturbance at the higher level.
The simplest example of this that I know of is also the first one I
fell for. Someone says "Hey, your sister just fell off the swing!" I
spin around and look, and see nothing of the kind. The other kid
screams out "Made-ya-look, made-ya-look!" and collapses laughing.
Once this has happened, it creates a conflict. What should I do the
next time someone tells me of some dire or interesting happening? If I
don't look I may fail to do something necessary, or miss something
interesting (it would have been interesting to see my sister fall off
a swing). The other has presented me with a problem (as you say). He
has not, however, presented me with a solution. It's up to me to
figure out that it doesn't really matter to me if someone else makes
me look by giving me false reports, not in comparison with seeing
something interesting or helping with a problem (I really would have
helped my sister). So the next time I'll fall for it cheerfully,
knowing that most of the time I can believe what people tell me
(except that kid, but even when he's the reporter I'll take the chance
because he might be telling the truth once in a while). From then on I
may look, but it's no problem. Not looking would cost too much. Let
him control the direction I'm looking for a few seconds -- what's the
big deal?
That kid, by the way, stopped doing that when someone else beat him up
for doing it.
The point is that I have altered which of your lower-level reference
signals you are (let us say) "needing" to use, given your high-level
reference signal which (as we have discussed before) can be said to
be >continuously operating throughout (with no disturbance prior to my
yelling "Fire!").
And my point is, "That's life." Any disturbance can do that. It
doesn't matter who is doing it or what is doing it. It's still not
control of anything that matters to me. In the final analysis, all
that I will NOT sacrifice in order to counteract a disturbance is my
system concepts.
Perhaps you will object that "alter" in the definition of
"controlling >another person" means "being the sole agent involved in
causing a >change in the variable."
Not at all; that's not what controlling means. I may tense my muscles
to keep from falling down when gravity is the sole agent involved in
causing a change in my posture. Gravity isn't controlling me; gravity
can't control anything. If, in addition, someone is pushing on me, I
will just shift the forces accordingly (by altering the appropriate
reference signals).
So alter must mean "being a contributory agent involved in causing a
change in the variable." That is, to alter a variable associated with
another person is to be necessary, but not necessarily sufficient,
for >a change in the variable.
This is getting positively Talmudic. Yes, I sort of agree with this.
Except that even if you do nothing, other causes may change the
variable. So "necessary" isn't in it. When you change your effect on
the variable, and allow that other influences may also be present, the
variable might do anything, including changing oppositely to your
influence.
Perhaps you will object that my (assumed successful) manipulation in
some sense REALLY didn't alter your control structure, only the way
it >was operating. That your reference signals weren't altered, only
which >were disturbed. OK, then if we take "which reference signals
are >disturbed" as a variable associated with you, I was controlling
that >variable, wasn't I?
No manipulation alters my control structure. It only alters signals.
All manipulations alter reference signals -- at the levels lower than
the place where the manipulation has its direct effect. The
manipulation is always resisted, successfully, at the level where its
direct effect occurs. The changes in reference signals that accomplish
the resistance take place at at lower levels. That causes the
manipulee no problems unless those lower-level changes alter some
other higher-level perception that is under control. Then direct
conflict will result (if youy insist on having an effect). If you yell
"FIRE" and can't show me where it is, I will put you under citizen's
arrest for public endangerment.
(2) I admit freely that the "short-term"/"long-term" nomenclature
isn't >optimal. I mean to suggest a distinction between the
controller's >altering the potential controlee's control structure
without/with >reorganization of the latter's control structure being
involved. "Long->term" alterations unquestionably involve changes in
the controlee's >reference signals, since such changes are the result
of reorganization.
You're using a very general term -- structure -- that isn't warranted
by the discussion, which so far has concerned only signals handled by
a structure of constant properties. All disturbances call forth
changes in (lower-level) reference signals. So do reorganizations, but
for a different reason. So with respect to whether reference levels
change, there is no substantial difference between long-term and
short-term. The main difference is in WHY they change.
And in "long-term" control of another's control structure, the
controller acts to control certain of his/her perceptions which will
be >controlled if the controlee's control structure reorganizes so as
to >"solve a problem" posed by the controller and which the controllee
wants to solve (shorthand for saying his/her higher-level reference
signal(s) require solution of the problem to be achieved).
Reorganization is not affected by higher-order reference signals, not
the way I model it. It's affected by intrinsic error, regardless of
its source.
If you attempt to control some aspect of my behavior by creating
circumstances that cause intrinsic error in me, you will certainly be
"influencing" my behavior. But you will not, with equal certainty, be
controlling my behavior. Even I can't predict what will result when I
reorganize -- how could you? I might end up solving the problem you
have posed, or I might end up solving a different one that makes your
puzzle irrelevant. Or I might just say "The hell with it -- you're
just making life difficult." There would be nothing you could do to
stop me -- without the use of force. But my point is that, as I said
above, while you can present the problem, you can't force what you
think of as the solution on me. I will find my own solution; in fact,
nobody else can. Only I can decide what constitutes a "solution."
If the problem is indeed solved thusly, then the controller (who
controls for his/her own perception that the controlee solves the
problem) can be said to have altered the controllee's reference
levels. >The control could involve direct force or threat, or not.
How many times do I have to scream "TO AFFECT IS NOT TO CONTROL?" It
is only to present a situation. All the hopeful controller can do is
present a situation by altering some variables. What the behaving
system will do about that, if anything, is completely out of the
controller's hands. All that the would-be controller can do is
disturb, usually without any concept of what is actually being
disturbed in the supposed controllee. A con man may get his money even
if the mark sees through the scam, perhaps because the mark feels
sorry for the con-man and doesn't mind handing over a few bucks to
give this pitiful creature a momentary sense of success. Haven't you
ever done that for a panhandler with a long sad unlikely story? The
con-man then thinks he is controlling one thing; in fact he is
affecting something else in the controllee and his sense of control is
a fantasy. [Notice, by the way, how it's OK to use sexist language
when you're saying something bad about someone. Nobody insists on
"con-person"].
So, such control is possible.
You're satisfied with your argument, but I'm not.
Its likelihood is another matter; its success requires that the
controlee not give up wanting to solve the "insoluble without
reorganization" problem and that he/she reorganize until a solution
is >found. Of course, the solution might be impossible altogether,
regardless of how much reorganizing occurs (remember Liddell's
"experimentally neurotic" animals who were given impossible
discrimination tasks?)
Now you're destroying your own argument. What is done about the
presented situation is out of the control of the putative controller.
So the putative controller is not a controller, but only an affecter,
only a disturbance. With respect to Liddell's rats, wasn't it lucky
for Liddell that so many rats volunteered to be in his experiment?
... but since the controller wants to see the controllee actually
solve >the problem, presumably the controller wouldn't pose an
impossible >problem except by mistake.
The controller can't make the controllee solve the problem, or predict
what solution will be found, or even predict that the controllee will
really try to solve the problem rather than just get rid of the pesky
controller. The controller isn't a controller; you can't make him one
just by calling him one. The controller is the problem, not the
solution.
Again, you might object that the new (reorganized) controlee's
control
structure is not determined solely by the controller.
It's not determined at all by the controller. Either you have sole
determination or you don't have determination at all. "Determined by"
means "unconditionally dependent on." If other influences can have an
effect on the outcome, actually altering it, then there is no control
or determination by anybody. If B is determined by A, then given event
A nothing but B can happen. "Joint determination," a phrase I vow
never to use again, is an oxymoron.
But will you object that the new structure is determined solely by
the
controlee?
No. Determination is not control, although control may entail
determination (as in, the reference signal determines the reference
level of the input signal). Control is varying an action so as to
maintain an input near a reference state despite disturbances. You're
shifting around among three different usages of the word "control,"
only one of which is consistent with PCT.
As I understand your theory of reorganization, a series of random
(pseudo- or not, and perhaps biased in certain ways -- that
possibility >doesn't matter to my argument) alterations in the control
structure >occur, and each is tested to see whether it solves the
problem (or >"conflict") which gave rise to the reorganization process
in the first >place.
Well, yes, but you have to be more specific. If there's a conflict in
the hierarchy, the conflict itself may be sufficient to start
reorganization because of the resulting error signals, but the real
reason may be that failure to control something (because of the
conflict) has led to a different intrinsic error -- being in conflict
about eating may cause your blood sugar level to fall. The criterion
for ending reorganization is not that the conflict be resolved,
because the reorganizing system knows nothing about conflict. It's
that the blood sugar, or the hierarchical error signal, be restored to
its proper state. It doesn't matter why it is restored. It could be
restored by accident, and reorganization would still cease. The "test"
for a proper solution has nothing to do with the solution. It has to
do with the intrinsic effect of finding any solution that corrects the
intrinsic error -- not a specific solution.
Finally ... if you object that what you really meant by not being
able >to control another person's reference signals is that you can't
alter >them reliably in a precise way which you have pre-specified, I
agree. >But you can control them as per the definition above of
"controlling >another person."
No, that's not my objection, although it's a good objection in that it
says you can't control unless you actually control. My objection is
that influences and effects are not control.
Perhaps we could save some time here if you would state the
conclusions toward which you are shaping all these arguments. I find
your uses of words very slippery, but they all seem to be aimed at
some final conclusion. Just what is this merging of the environmental
and control views that you want to accomplish? If it takes this
elaborate and convoluted an argument to support the idea that the
environment (or other people) actually does have some controlling
effect on organisms, I can't help wondering which came first: the
conclusion or the reasoning.
I'm with Rick on this subject: as far as I can tell, I'm just making
straightforward deductions from the principles of PCT. Any "ideology"
that's involved follows from these deductions. I don't see anything
straightforward about your deductions; you keep switching around among
"affect," "influence," "determine", and "control" as if these were
equivalent terms. You can't prove control by demonstrating influence
or any of the other terms. Physical environments can't control
anything, lacking as they do the required organization. Other people
can't control a person's behavior except in a trivial sense; they can
control (non-coercively) only variables of lower level than the
variable they disturb. As far as I can tell, these results follow
strictly from the theory, and need no ideology to sustain them.
So what's your ideology?
···
----------------------------------------------------------------------
Best,
Bill P.