From Greg Williams (920830 - 2)
Bill Powers (920830.1545)
Control systems don't control their actions; they control their perceptions.
By applying a disturbance to a controlled variable and thus disturbing, or
attempting to disturb, the perception of it, an outside party can cause the
action of the control system to vary. It will vary, predictably, so as to
remain equal and opposite to the disturbance. That is why the disturbance has
essentially no effect. So the outside party can control the amount and
direction of the action that the control system is producing.
More precisely, it seems to me, the outside party can control his/her
perception of the amount and direction of the action produced by the control
system. Right? (An aside: Of course, a disturbance can have TREMENDOUS effects
on the control system's outputs -- but not (assuming good control) on the
outcomes. I think it can be quite misleading to leave the word "effect"
unqualified; in particular, it can lead one to think (falsely) that an
alteration in an UNcontrolled variable has "essentially no effect" in general,
when it in fact can result in switching of control to entirely different
percepts than were being controlled before the alteration.)
Controlling "a person" means controlling the value of some variable
associated with that person. Some of those variables (the amounts and kinds
of actions) are controllable if the controlled variable is known. Others
(perceptions, controlled variables) are not.
OK, now I think I understand your definition of "controlling another person":
it is controlling one's perception of certain actions produced by another
person. Correct? If so, I see only one problem in replacing my "manipulation"
with your "controlling another person." (See my next comment below.) My
confusion about what you meant was due to the discussion on the net a few
months ago wherein most everyone agreed that one cannot "causally" PCT-control
others' controlled perceptions, as you explicitly note above, simply because
the would-be controller does not have direct access to others' reference
signals, which causally determine their (the others') controlled perceptions.
If you look back at my "Prolegomenon," you'll see that I was trying to be
careful to deny that one can alter the current reference-signal structure of
another (but it is an altogether different question whether one can, by
altering another's perceptions -- those which CAN be altered, of course --
play a role which contributes to a different FUTURE reference-signal structure
than would otherwise exist; but that is at a tangent to my present concerns).
If the manipulee is aware of the manipulator's goal, the manipulee can
simply adopt it and control for it. In that case the manipulator has
no control at all.
If the manipulee adopts and controls for the manipulator's goal, the
manipulator doesn't necessarily stop controlling for that goal. Why should the
manipulator stop? I suspect that usually BOTH manipulee and manipulator would
continue controlling (for their respective perceptions of achieving the
manipulator's goal). And the upshot, then, would be that the manipulator has
an easy time of controlling, rather than no control at all. To say that the
manipulator would have no control in this case seems to me to require an
addition to your definition of "controlling another person," namely the
requirement that such control cannot result from the controllee adopting the
controller's goal overtly. I had no such requirement in my definition of
manipulation, but I had another requirement, namely that the manipulator's
altering some of the manipulee's perceptions helped to determine the
manipulee's control processes which resulted in certain of the manipulator's
perceptions matching their reference signals. So, in the case you describe
above, given our respective definitions (yours not amended), it appears that
the influencer (I can't say manipulator) has control but is not manipulating.
So our definitions part company, right?
... what I was trying to get at was manipulation as making the
influencee control (to SATISFY his/her interests or wishes) in such a
way as to ALSO satisfy certain interests/wishes of the influencer.
How can you MAKE the influencee control for anything?
Gloss it as "provide the opportunity for" control of perceptions which would
not have been controlled for in the absence of the influencee's perceptions
being altered by the influencer.
If controlling for something satisfies the influencee's own higher goals,
that will be done. Otherwise it won't. The influencer can suggest possible
goals, perhaps even goals the influencee is glad to be told about. But
nothing at all will happen as a result unless the influencee translates the
verbal suggestion or the seen example into a personal perceptual reference
signal and activates it. Only a higher system in the influencee can do that.
Yes, that's what I say, too. In manipulation, control of certain perceptions
by the influencee results in control of certain perceptions by the influencer.
The influencer wouldn't see what he/she wants unless the influencee were
controlling certain perceptions (possibly with no knowledge of the
influencer's desired perceptions). But for it to be manipulation, WHAT the
influencee is controlling for must be partly the result of actions of the
influencer. Now, of course, I don't mean to say that the influencer alters the
reference-signal structure of the influencee, but rather that the influencer,
by altering "don't-care" perceptions of the influencee, provides the
opportunity for the influencee to control perceptions which the influencee
would otherwise not have controlled. To repeat an example, "Look at this
wallet I found here." No wallet, no control for taking the money in it. But
the potential (so to speak) for control for taking the money was already there
in that "good" mark who has shown "greediness" (not so subtly!) to the con
man. So the con man has to work with what he finds in the mark's control
structure, but he alters the mark's perceptions in ways so as to allow the
expression of control by higher systems in the mark to benefit him, the con
man.
What an influencer or manipulator wants is to be the AGENT that takes credit
for the result. In order for that to happen, the influencee must NOT be the
agent.
By my definition, all the manipulator wants is for certain of his/her
perceptions to match his/her reference signals. The perceptions need not
include that he/she is "the AGENT that takes credit for the result." The
influencer, by my definition, doesn't necessarily want ANYTHING -- influence
can be "accidental." Con men know full well -- and say it -- that the mark
gets the credit for his/her actions. Even the best con men repeat the old
adage that "you can't cheat an honest man [or, presumably, woman]." (Actually,
the old adage doesn't mean that one can't persuade an honest man to part with
some of his resources by deceptive means; it means that it is best to con
marks into acting dishonestly so they won't tend to run to the police when
they realize -- if they ever do, some never do -- they've been "had.")
Unless your kids have some reason to want Pat NOT to see them eating food she
considers healthful, they don't care what she's controlling for by setting
various kinds of food on the table as long as they like the result. Pat is not
controlling them, or even disturbing or influencing them. Their controlled
variables are undisturbed.
By my definitions, she is CERTAINLY influencing them. Without Pat, there would
be no tasty noodles. She alters their perceptions. Further, if she provided
tasty "junk" (her definition) food instead of tasty "healthy" (her definition)
food, they would gobble the "junk" food, and she would not be able to control
her perception of seeing them eat "healthy" food. The situation certainly
seems to meet the criteria of my definition of manipulation. The kids'
controlled variables are undisturbed, yes, as I've been saying all along.
WHICH percepts they are controlling (with their UNDISTURBED reference levels)
are different because of Pat. And Pat's goals are met because (in part) of
what Pat did; had she done otherwise (as suggested above), her goals would not
have been met.
They are the agents of their own actions. Whatever consequences of eating one
kind of food rather than another Pat may imagine to be going on inside the
kids, nothing is happening inside them that disturbs their goals one way or
the other -- or their intrinsic variables either, one way or the other.
Yes, I agree. They are, in my terms, manipulated agents of their own actions.
In my Random House dictionary, the first meaning [of "manipulation"] is
simply related to skill. The second meaning is "to manage or influence by
artful skill: _to manipulate a person_."
That certainly sounds like what I'm trying to investigate, using PCT ideas. No
connotation of negativeness; skill is needed; influence is presumed -- it's
all there, waiting for a theoretical underpinning.
What other kind of control is there but control in the PCT sense? If
the manipulator wants to have a certain effect on the manipulee, and
carries out actions that have that effect even despite disturbance,
isn't that just plain garden-variety control?
I think I'm straight on this now. Thanks.
If the control is not against a person's will, then it is control of
something that person considers unimportant, like the position of the finger
in the rubber-band game. Of course at some point the position of the finger
may become important to the manipulee, as when the finger approaches a hot
spot. The control by the manipulator will then abruptly be lost.
With artful skill, some manipulators (or, if you don't like the word, simply
"con men") can (almost literally) burn fingers without the manipulation
ending. I'm still serious about understanding the nature and limits of this
very real phenomenon, which I think is ubiquitous in family interactions,
business dealings, education, therapy of all sorts, entertainment, religion,
politics (robustly, as they say), etc., etc. I welcome additional suggestions.
The literature on manipulation is surprisingly scanty, and mainly empirical.
Best,
Greg