A comment on PCT correctness

CSGnet Archive CSG1995
1

****** FROM CHUCK TUCKER 951219.10:02 EST ******

          I realize that I frequently begin my post with a
          disclaimer that "a more extensive statement will
          follow" and I only rarely comply with my own
          instructions and promise to you but I am doing it
          again in this post. I was struck by Rick's statement
          (951218.0900) on the "leaving of Ed Ford" that "The
          ultimate arbiter of 'correctness' on CSG-L should be the
          degree of match between the PCT model and real (his
          word!) behavior." [my comments enclosed in ()] This
          statement was given an "Amen" by Martin (951218 12:30)
          but not mentioned at all by Bill's (951218.1715 MST) post.
          If this is so, why is this criteria rarely if even used
          on the net? Is it like my failure to comply with my
          own instructions to myself or is there no intention by
          the major discussants on CSG-L to comply with this
          criteria? Whatever the answer I find little evidence
          on the net of using the criteria.

          When THE TEST (TT) was mentioned several months ago (so it
          seems to me) there was an extensive argument about what TT
          is, what TT means, whether it can be used with human beings
          but never any report of TT used with "real behavior." When
          I posted the only description (that I am aware) of TT on the
          net it was only used by Martin and totally ignored by the other
          discussants EXCEPT IN ANSWER TO Martin. For me, the
          discussion "ended" with a request for a proposal to do some
          "research" with a human being and a computer program (which
          should BTW should be right up Martin's alley since that is one
          of his major interests). No one has yet designed any
          research on the net that involves "real" human beings engaging
          in "real" behavior with each other.

          The posts about "science as mush or non-mush," "whether
          reinforcement is a fact," "whether RT=SR," "whether RT
          actually explains better than PCT or vice versa," although
          very interesting [I ENJOY READING THESE POSTS AS LITERATURE,
          FOR THE STRUCTURE OF ARGUMENTATION DISPLAYED THEREIN, AND
          FOR SOME INFORMATION] but not one report of "real" behavior
          and how it might "fit" with the PCT Model did I find therein.
          Nor did I even find any reports (a la "case study" research)
          of any person's experience with TT or "fitting" said
          experience with the PCT model. What I did find was the
          phrase "behavior is control," "PCT is about control,"
          "conventional research is irrelevant to PCT," and "if
          we consider it mathematically," repeated over and over
          as if some how such behavior would confirm the "fit"
          between "real" behavior and the PCT model. I'm sorry
          but for me it does not compute.

          I strongly recommend that we begin to report on the net
          the results of "real" research with "real" human beings
          that deal with "real" behavior to see if there is a "fit"
          between "it" and the PCT model.

          As a contribution I will report on some of my recent
          research with "self system concepts" on this net early
          next year.

          Regards, Chuck

2

[Martin Taylor 951219 12:00]

CHUCK TUCKER 951219.10:02

          As a contribution I will report on some of my recent
          research with "self system concepts" on this net early
          next year.

I think that the special issue of IJHCS will include an updated and combined
version of Robertson and Goldstein's two (rejected) papers on this issue.
You might want to ask Dick if he would like you to review it for him.

The issue will also include Rick's work with the satellite controllers, my
own prototype development of a tool for designing human-computer interfaces,
and one from a Dutchman called Reinder Haakma on the design and testing of
a real interface to a commercial digital tape recorder. There will, of course,
also be the theoretical papers, and I'm still hoping to get "Models and
their worlds" from Bill P and Tom Bourbon.

          When THE TEST (TT) was mentioned several months ago (so it
          seems to me) there was an extensive argument about what TT
          is, what TT means, whether it can be used with human beings
          but never any report of TT used with "real behavior."

As for "the Test," I see it more as a way of assessing whether a proposed
function is something like a controlled variable, whereas in the real
world the problem is often more of exploration, looking for potentially
controlled or controllable variables. The signs of conflict, for instance,
are evident to a third party, but it is seldom evident which correlated CEVs
have given rise to the conflict. It is even less evident when the conflict
is between coordinated groups of people, as it is in a war. When there is
no evident conflict, I think it is even harder to determine what "real" people
are controlling for, and in a practical sense that's what matters because
that's what may lead to future conflict.

                                                         For me, the
          discussion "ended" with a request for a proposal to do some
          "research" with a human being and a computer program (which
          should BTW should be right up Martin's alley since that is one
          of his major interests).

Yes, but I'm a poor and slow programmer now living on a pension (as I presume
Bill Powers is), with no support such as students or money to run any
experiments that might be proposed. But I have produced a few starts
at such experiments. One I call "Revere" (after Paul Revere) deals with the
perception and control of uncertainty, but it doesn't work very well.

I think the power of PCT is in its direct applicability to so many real
world problems. So many puzzles make sense when seen as issues in
perceptual control, and so many political policies are seen to be stupid
and self-defeating because they attempt to force action rather than
disturb perception. But it seems to me that there is little hope in
demonstrating to the public the nature of the difficulty, when so many
(what is it, 30%?) cannot even find the Pacific Ocean on a world globe.
Those are the voters who believe in the self-defeating policies put
forward by confident-sounding politicians.

Anyway, I agree with you in spirit, but I'm rather dispirited about the
possibility of doing anything useful.

Martin