Sociocybernetics??? Tom's Study

CSGnet Archive CSG1993
1

[From Rick Marken (930513.0900)

Chuck Tucker --

STATEMENTS FROM A SOCIOCYBERNETIC PERSPECTIVE OF HUMAN CONDUCT*

This is a curious set of sentences. Why did you post them? Are they
examples of strings of words that someone might create if they
were only vaguely familiar with PCT? Some of these statements
seems patently wrong given the meanings that I get from words
like "action". For example:

People are responsible for their action, but <<not>> for everything
that happens to them.

People (as reference signal sources) are responsible for (cause)
their perceptions, NOT their actions; actions are caused by
disturbances to the perceptual variables that are controlled.
Thus people are ONLY responsible for what happens to them; not
for the actions that are required to keep it happening. While it is
true that people are not responsible for everything that happens
to them, this is not a particularly significant observation. If people
are not responsible for it, then it is an uncontrolled aspect of what
happens to them -- so it is irelevant to them.

People guide their actions by directions they give themselves.

Disturbances guide actions; people don't guide their actions at all.
This is one of the most fundemental and important facts about
control system operation. People guide their perceptions (what
happens to them), not their actions.

Most of the other statements are just too ambiguous; I can find
ways to understand them so that the meaning seems consistent with a
quantitative understanding of the behavior of a hierarchy of
control systems; but they often evoke understandings that strike
me as being quite wrong.

* Slight modifications of statements used in the courses of Bob Stewart.

I'd tell Bob Stewart to save the sentences for church and to spend the rest
of the week studying with the real thing -- a model that controls.

Tom Bourbon (930512.0123) --

Your description of the two-person cooperation experiment was excellent!

I just want to point out some ways of looking at what's going on that might
help people see that there is, indeed, quite a bit more here than "stick
wiggling" -- there is a real, complex social interaction going on.

First, it is interesting that, in order for either subject to control any
pattern of the three lines, BOTH must control a lower order variable --
the difference between two lines; the left subject must control cr-cm
and the right subject must control cl-cm. But NEITHER subject can control
the appropriate difference unless the other is also controlling. So this
is an intrinsically social task -- both subjects must control or neither
can control. Sounds like a social phenomenon to me -- likemoving a table;
both people must lift their ends, simultaneously, or neither accomplishes
the goal of moving the table.

Second, it is interesting that success depends on the subjects learning what
variable to perceive (and control) not on how to act. The left subject must
learn that cr-cm is the perception to control; the right subject must learn
that cl-cm must be controlled. The subjects must hit on this nearly simul-
taneously because, unless one subject is controlling his variable, the other
subject cannot control hers. This is what makes learning to control the
higher order variable (pattern of the three lines) so hard in this task. The
differences to be controlled may seem like a simple perceptions to learn, but
it is hard because control of that these perceptions depend on the other
subject's efforts to control as well.

Third, the difference between selecting the correct perceptions to control and
the incorrect ones is the difference between cooperation and conflict. If the
left subject decides to control cl-cm instead of cr-cm, he is putting himself
into conflict with the other subject (possibly). This is the most fascinating
and important part of Tom's experiment for me: the difference between
cooperation and conflict in this experiment is a simple difference of
perception. If, while cooperation is occuring, the left subject does nothing
more than "change his mind" and start to control cl-cm instead of
cr-cm there is suddenly VIOLENT conflict -- both subjects moving their
handles wildly in order to control their perceptions -- and neither achieving
their higher order goal (of having the lines all in a row, for example).
The conflict requires no direct physical contact, no strategy for action that
conflicts with the other person's strategy, etc. The difference beteen
conflict
and cooperation sits on the razor's edge of perception; by just "looking" at
the situation slightly differently (trying to control a slightly different
aspect
of the situation) we get fruitless conflict instead of a harmonious dance.

I think this is a brilliant study, Tom. The "real" social scientists might
not
be interested but your fellow pseudoscientists and rubber banders think
it's terrific.

Best

Rick

2

FROM CHUCK TUCKER 930513.14:52

        I find it quite curious that you can find very little of
        of meaning that you can make of the statements I have put
        on the net (this is about the third time I have put them
        there by the way) yet Bill seems to understand them quite
        well - in fact (a word I know you like but understand
        differently than I do) I agree with all of his interpreta-
        tions of my statements in PCT terms. I suggest that you
        read his post of today for a restatement of my statements
        in PCT terms.

        I posted the list (as I noted) for the statements about science
        and laws but I will still stand by the other statements. I
        do guide my actions by the directions that I give myself by
        setting reference signals and acting to accomplish them. I
        increase the liklihood of doing what I tell myself to do by
        making arrangements (other actions based on reference signals)
        that would reduce disturbances knowing that I will still have
        to deal with whatever disturbances I encounter. I tell myself
        "Hold your arm straight out from your body so that the angle
        of the arm and the body is 90 degrees." Then I proceed to
        move my arm to achieve that reference signal; I am controlling
        my action (behavior, arm movement) to accomplish that goal. If
        that does not fit in your model then your model is wrong but you
        are not stupid; just wrong.

        I commented earlier on the notion of "understanding" and have thought
        about it some more since I have been making judgements about students
        as to, in my judgement, if they have understood what I have asked them
        to read and what I have said to them this semester. I think now that a
        use of "understanding" as you have said that you are trying to do (i.e,
        people should understand PCT before they use it; I am not trying to
        convince people just trying to make them understand) is a way of forcing
        another to act to your specifications (as for example, use particular
        words in particular ways). It is a vary sly way of doing it but it is
        an attempt to force another's action.I would suggest that you might want
        to rethink this matter and make it more consistent with your model of li
        or your model of "boss reality" (whichever you wish to use today).

        On BOSS REALITY. I just find it unnecessary to deal with the problems th
        interest me to have both a model of "Boss Reality" and a model of the mo
        of the "boss reality." If I ever have any reason to refer (people refer
        with words; words don't refer) to what you seem to be modelling as "boss
        reality" I simply say "a world that is there." Since I can't experienc
        "it" I have nothing useful or necessary to say or write about "it." If
        you want to continue to do science with the conventions and standards
        used by the neorealistic postitive empiricists then you will have to
        do it and understand it they way they do. I don't care to be disturbed
        by such silly nonsense.

        I will give a copy of your post to Bob Stewart but I doubt that he will
        more that raise an eyebrow and mudder something like "Boy, has that guy
        a lot to learn." If he does otherwise I will report it in all of its
        empirical spendor to you.

        Regards,

                 Chuck

        PS I am writing on a MAC and not familiar with it so I noticed that some
           letters where "cut off" at the end of sentences; sorry for the
           disturbances - see all of my arrangements did not get rid of them.

3

From Tom Bourbon (930513.1611)

                     FROM CHUCK TUCKER 930513.14:52

Just a quick question, Chuck, about the following paragraph.

       I posted the list (as I noted) for the statements about science
       and laws but I will still stand by the other statements. I
       do guide my actions by the directions that I give myself by
       setting reference signals and acting to accomplish them. I
       increase the liklihood of doing what I tell myself to do by
       making arrangements (other actions based on reference signals)
       that would reduce disturbances knowing that I will still have
       to deal with whatever disturbances I encounter. I tell myself
       "Hold your arm straight out from your body so that the angle
       of the arm and the body is 90 degrees." Then I proceed to
       move my arm to achieve that reference signal; I am controlling
       my action (behavior, arm movement) to accomplish that goal. If
       that does not fit in your model then your model is wrong but you
       are not stupid; just wrong.

Apparently I am wrong along with Rick. As just one example, in Models and
Their Worlds, I (along with Bill, who must also be wrong) tried to test the
results were a system to act so as to control its actions. It didn't work
out very well. That was the cognitive or plan-driven model, and it failed
in the greatest number of circumstances (it succeeded in the fewest
circumstances).

I thought living control systems controlled their perceptions, not their
actions. That is certainly the way it seems to work best in the models I
have seen and worked with, but I am willing to be shown I am wrong.
How should I change the plan-driven model so that it might succeed, and
the PCT model changed so that I can eliminate the wrong idea that it
controls its perceptual signals? A change like that would certainly make a
big difference in all of my interactive programs, and in CROWD, ARM, E-COLI,
and many other PCT modeling programs.
Until later,
  Tom Bourbon

4

From Tom Bourbon (930514.1003)

Chuck, in your post (FROM CHUCK TUCKER 930513.14:52) you wrote, in reply to
Rick:

       I posted the list (as I noted) for the statements about science
       and laws but I will still stand by the other statements. I
       do guide my actions by the directions that I give myself by
       setting reference signals and acting to accomplish them. I
       increase the liklihood of doing what I tell myself to do by
       making arrangements (other actions based on reference signals)
       that would reduce disturbances knowing that I will still have
       to deal with whatever disturbances I encounter. I tell myself
       "Hold your arm straight out from your body so that the angle
       of the arm and the body is 90 degrees." Then I proceed to
       move my arm to achieve that reference signal; I am controlling
       my action (behavior, arm movement) to accomplish that goal. If
       that does not fit in your model then your model is wrong but you
       are not stupid; just wrong.

Yesterday (Tom Bourbon 930514.1000), I posted this request:

Apparently I am wrong along with Rick. As just one example, in Models and
Their Worlds, I (along with Bill, who must also be wrong) tried to test the
results were a system to act so as to control its actions. It didn't work
out very well. That was the cognitive or plan-driven model, and it failed
in the greatest number of circumstances (it succeeded in the fewest
circumstances).

I thought living control systems controlled their perceptions, not their
actions. That is certainly the way it seems to work best in the models I
have seen and worked with, but I am willing to be shown I am wrong.
How should I change the plan-driven model so that it might succeed, and
the PCT model changed so that I can eliminate the wrong idea that it
controls its perceptual signals? A change like that would certainly make a
big difference in all of my interactive programs, and in CROWD, ARM, E-COLI,
and many other PCT modeling programs.

I would still welcome your suggestions on how I should revise the two
models, but last evening I realized another person should be
invited to participate in the revisions. In your reply to Rick, you said:

I will give a copy of your post to Bob Stewart but I doubt that he will
more that raise an eyebrow and mudder something like "Boy, has that guy
a lot to learn." If he does otherwise I will report it in all of its
empirical spendor to you.

I had forgotten that the list you posted originated with Bob Stewart. If he
is unfamiliar with Models and Their Worlds, perhaps you can guide him
through it, then let us know his ideas.
Until later,
  Tom Bourbon