From Greg Williams (920918 - 3)
Bill Powers (920918.0800)
I'm sorry about the harsh tone of my last post. I was frustrated. I
really like you and Pat a lot, and I always will.
I understand. I was frustrated, too. Perhaps we would be less so in this
"lovely argument" -- as you termed it in a private post just a couple of days
back -- if we each don't frame it as trying to convince the other of anything,
but as trying to see how far we can go regarding control-theory implications
for social interactions. So, back to the lovely argument....
"Purposive influence" is an oxymoron if it doesn't mean simply
"control." If you have only an influence on what someone else does,
and not control, then you can't predict what the other will do, and
you can't vary what you do in order to make the other do any
particular thing. In that case it can't be a purposive influence.
I agree. Explicitly, influence, if "purposeful," is control.
An "influence" is simply one among many independent variables. It does not
have a determining effect on the result. Over the long run the effect of an
influence can be detected only statistically, by averaging out the effect of
all the other independent variables that act at the same time, also
influencing the same outcome.
OK, a non-purposeful influence (i.e., not control) is as you say. And I'll add
that the non-living portion of an organism's environment (but, now that I
think of it, I don't know whether that includes at least some artifactual non-
living feedback control systems -- but that's a detail to be pondered later)
can ONLY influence organisms and CANNOT control ANYTHING, simply because that
portion of the environment has no reference signals.
To convert a mere influence into control, you have to monitor the
influenced effect, compare it with a reference level, and adjust your
influenceS whenever there is a deviation from the effect you desire.
Yes.
Now you compensate for variations in all the other independent
variables, not by knowing what they are but simply by monitoring the
final effect you want. If you can't monitor the effect you want to
influence, and vary your actions to oppose changes in it, then you
have only influence, which is necessarily statistical, and not
control.
I agree.
In an example such as feeding children healthful food, you may be
influencing their health, but you're not controlling it.
That is correct. What Pat is doing is controlling her perception that the kids
eat food which she THINKS is "healthful," according to her definition of the
term.
You also may not be influencing their health.
Right. In controlling her perception that the kids eat food which she
considers "healthy," the food she provides might not alter their actual health
relative to their health if Pat didn't control that perception.
In principle, according to statistical theories of nutrition, "people" are
better off eating some foods than others. If your children are healthy,
however, there is no way to tell whether that is the result of what they eat,
or the effect of powerful biochemical control systems inside them that can
convert a wide variety of inputs into sufficiently useful compounds and reject
a wide variety of useless or deleterious inputs without harm, or simply a
result of the fact that the food you feed them is as good as any other kind.
You said it far better than I did above.
At best you can say you're not making them unhealthy, for the simple reason
that they're healthy and therefore NOTHING is making them unhealthy.
In fact, I think, Pat actually COULD be making them unhealthy, but "luckily,"
something else is compensating for that. She really has no way to monitor all
of the sources of the kids' health unless -- MAYBE -- she could keep them
penned up some way.
Unless you're continuously monitoring their biochemistry and correlating it
against variations in their diet, you have no knowledge of the actual effects
of their diet. If you feed them only one kind of diet, you have satisfied one
of your own reference levels, but have no information on the effects of this
diet as opposed to others on your children. You are certainly not controlling
their state of health. You won't know whether you're even controlling what
they eat until they demand to eat something else. Then you'll find out.
Right. Pat will know BY DEMONSTRATION rather than just BY GUESS whether she is
controlling for her perception that the kids are eating what she considers
"healthy" only if there is an error signal in her perception, which she then
able to correct.
You've cited your nutritional customs several times as an illustration
of purposive influence of another. The only purposive influence in
this case is on your own perceptions of the effects you hope to have
on your children, in relation to your own beliefs about nutrition.
Being careful here, I would want to rephrase this and say that Pat GUESSES
that she is controlling her own perception of the kids' eating food she
considers healthy if she HAS NOT SEEN an error signal in that perception and
managed to correct it, but that she has WITNESSED successful control of her
own perception of the kids' eating food she considers healthy only when she
HAS SEEN an error signal in that perception and managed to correct it.
However, assuming that Pat IS successfully controlling her perception of the
kids' eating food she considers healthy, she is accomplishing that control by
INFLUENCING the kids. Possibly you don't buy this last claim. If so, please
explain your problem with it.
Those perceptions you can control, because they're your own.
Yes.
Those effects you can't actually control, you can imagine.
At least SOME effects one can't control can be imagined -- there might be some
such effects one can't control because one is ignorant or stupid in certain
ways.
There is no problem with making sure your children eat the things you
want to see them eating, because they are too young to find their own
food and they can't leave to find a wider variety from which to choose
for themselves. You have physical control over them, as your parents
had over you.
There WOULD be a problem if Pat gave our kids food which she considers healthy
AND that tastes bad to the kids. Only because they WANT to eat good-tasting
(to them) food AND because Pat gives them good-tasting (to them) food is there
no problem. Pat has made (occasional) mistakes about whether particular kinds
of food are tasty to one or the other kid, and, believe me, then there are
PROBLEMS, regardless of our having (some) physical control over them and it
being inconvenient for them to get food independently of Pat. If the food
tastes VERY bad to them, they'll just not eat, period.
So the crux of what Pat must do to control her perception of the kids' eating
food she considers healthy is to give them food which is BOTH "healthy" (to
her) AND "tasty" (to the kids). It is her purposive COMBINING of the two food-
properties which allows her control of her perception to be successful.
But you have another advantage, which is that you perceive in ways and
at levels where your children are not yet competent and don't feel
competent. As long as your children don't have their own goals for
what they eat, they will simply accept what they are given and believe
what they are told about it.
That is correct. If the kids adopt the goal of, say, eating not-very-tasty
"cool" food (raw fish soaked in horseradish? who knows??) to impress their
peers in preference to eating tasty food, Pat would be well advised to start
dishing up "cool" AND "healthy" (to her) food -- if such food exists. And if
the kids decided to go on a fast, there is simply NO food she could serve to
control for her perception of them eating "healthy" (to her) food.
When you're supplying information, activities, attitudes, and so on at
the level where it has an effect, you are the only control system
involved at that level. You present a world exactly in the form you
want, which the child is only dimly beginning to perceive.
I don't follow you here. Can you expand on this? It seems to me that Pat is
NOT presenting the world "in the form she wants" to the kids, but tacitly
acknowledging their wants in order to successfully control her perceptions
which she wants to control. No way can she disregard that they want a certain
kind of food (tasty, to date). The "trick" of her achieving control of her
desired perception is in her pairing WHAT SHE WANTS with WHAT THE KIDS WANT.
You can't control what the child is going to acquire from this new world; it
won't be exactly what you think, and it will include many things you were not
aware of presenting. The child will often take a disliking to something you're
trying to get across, and there's nothing you can do about it, short of
force.
Sure. Through time, the necessary "pairing" as per above will alter in ways
Pat cannot predict. That's where using The Test comes in. When her attempts at
control of her perceptions (which depends for success on having a good model
for what the kids want at any time) fail, she will, for example, ask the kids
what kind of food they like NOW. And then try giving them food which is both
"healthy" (to her) and is now liked by them. Of course, her model could be
wrong (perhaps the kids lied about what they like) -- and she tries again to
use The Test, perhaps this time by not asking the kids, but simply observing
them in various situations.
Parents can purposively influence their children as long as they don't
interfere with levels of control that have already developed or appeal
to levels that are still inconceivable to the child.
Yes. They also must take into account (as per the "pairing" discussed above)
what the kids want. Actually, that goes for purposeful influence in general.
You will realize that your children learned a lot more from how you WERE than
what you DID, and that the impetus for that learning came from them.
And how -- Bateson's "deuterolearning" is a prime source of embarrassment for
parents, years later!!!!
Not much argument in this post, I think... but I hope you still find it
lovely.
ยทยทยท
-----
Two entities [Bill and his cat], not controlling each other but acting to
hold up two ends of something that can't be carried by one end, and each of
them knowing it and counting on the other. Doing something together that
neither one could do alone.
Symmetrical (more-or-less) purposive influence?
----
Let's by all means explore the new TUTsim. But let's also lay a snare for
Barry Richmond and Stella, because that brings in the Mac world. Of course
you're perfectly free to say you'd rather develop your own system...
All three possibilities sound interesting to Pat and me at this stage. If any
IBMers want more details on the current version of TUTsim, which starts at
$130 -- much less in student versions which are limited in various ways -- I'd
be happy to provide ordering details.
I suspect that it wouldn't take much to convince the STELLA folks to add some
bells and whistles specifically desired by PCTers.
Optimistic again,
Greg