[From Bill Powers (2000.02.29.1041 MST)]
Paul Stokes (attachment on 2000.02.28)
The problem with your arguments concerning collective perceptions is that
you seem to confuse agreeing about what words to use with sharing a common
experience to which the words refer. In your example of the university, you
describe a few ways in which professors and students act (assuming that the
images I get from those words are the same ones you get), and say that any
normal person would understand that this is a "university." I can agree
that most people might use that word, but I still don't agree that they
would necessarily have the same perception of that to which the word
"university" points.
Your whole thesis seems contained in the fourth paragraph of your paper:
Perceptions and collective perceptions are part of the taken for granted
world of everyday life. They form the foundations on which the edifice of
common sense is erected.
I completely agree with that, yet I claim that people do not in general
know what other people mean by words, or how the world appears to other
people. As you say (paraphrased by me), they take it for granted that other
people experience a world just like theirs, but I claim that if we were to
do careful tests, judging not by the words people use but by the variables
they control, we would find that there is far less agreement than people
assume there is.
Of course you are quite right in saying that a lot of reality-testing goes
on, and that major disrepancies tend to be discovered because of conflicts
they produce. But simply eliminating conflict is not sufficient to prove
that one perception is like another, particularly at the higher levels of
organization. I think it is something of a miracle that we can communicate
as well as we do, or think we do. Yet even with all the communication that
goes on, getting an idea across to someone else is a difficult and chancy
business -- and I speak as something of an expert on that subject. People
may want to understand, but too often "understanding" means fitting a new
idea into their old ones, and they quit checking up on their understanding
when they get the pleasant feeling that they grasp what has been said. For
too many, this means being satisfied that nothing they knew before has to
be changed or abandoned.
I think there is another and more useful way to understand collective
perceptions -- Rick Marken mentioned it. Kent McClelland was also trying to
make sense of this notion, and particularly the notion of virtual reference
levels (which presupposes, to some degree, common perceptions).
In any group of people there will be some variables they all believe the
group is trying to control, particular _social_ variables such as
cooperation. There may, indeed, be some objective similarities in the
controlled variables as different people perceive them, and the reference
levels may, indeed, be somewhat similar -- at least in the respect of
setting them "high" as opposed to "low." But there will inevitably be
differences, both in the nature of the perception each person is trying to
control, and in the exact degree of that perception that is desired. These
differences will at times result in tension within the group, but if they
don't amount to outright conflict, the group can live with them.
We could say that a _collective_ perception is simply some sort of vector
average over all the perceptions that the individuals contain, and that the
_collective_ reference level is a similar average over all the desired
states (considered individually, each person having a desired state for his
own version of the supposedly common perception). From this we could see
the collection of people as a single _virtual_ control system controlling a
single _virtual_ perception relative to a single _virtual_ reference level
(for each dimension of the social interactions that we can define).
This concept can work when the individuals in the group are not in conflict
to the degree that none of them can act effectively. The virtual perception
being controlled by the group control action may not fit any individual's
perceptions perfectly, but each individual will perceive his own version of
the perception as being present to _some_ degree. And the virtual reference
level may not be quite right for any individual, but it may not depart too
much, operationally, from the state each individual wants for his own
perception. Thus all of the individuals in the group may end up feeling
some degree of error, and pushing to some degree to correct it, but if the
social interaction is successful the whole ensemble can be in an
equilibrium state, no person being too dissatisfied.
When an external disturbance occurs that tends to change the virtual
controlled variable, it will push that collective variable away from its
equilibrium state. It may help move the situation closer to a few people's
reference levels, but because the prior situation was in equilibrium, it
will cause greater errors in more people, and there will be a collective
error signal. That error signal will lead to actions by each individual to
correct his or her own _actual_ error signal (varying in size from one
person to another), and that will lead to a set of actions all of which
have a component, at least, opposed to the disturbance. So the result is
opposition to the disturbance by the _collective_ or _virtual_ control system.
This picture does not require either perceptions or reference levels to be
identical in all members of the social unit. All it requires is some degree
of similarity. I think this is a far more realistic idea than to claim that
different people can somehow "share" perceptions or reference levels. The
latter concept assumes that there is just one perception and one reference
level to be "shared", rather than a large collection of perceptions and
reference levels that are somewhat similar across a group.
Best,
Bill P.