Bill's Journey

[From Rick Marken (939401.1500)]

Bill Powers (930401.0800) --

Having wandered afield from my original questions, I suddenly
think I see an answer to the paradox, and our disagreements about
information theory:

Wow. This one's even better than the one about information being
in uncontrolled perceptions.

I think we can find that source of information: it is in the long
process by which the control system becomes organized.

The principles of control theory are
satisfied by the given form of the output function (and other
functions in the system) and the current states of all the
variables. The principles of information theory are satisfied,
without conflicting with the principles of control, by showing
that the information missing from the present form of the
perceptual signal is being supplied by reorganization based on
the remaining information in the perceptual signal and stored in
the form of the output function itself.

This is really lovely!

Only one problem. The IT folks might notice that this solution
means accepting that there is no information in the perceptual
signal that is actually used as part of the process of control
(that is, that is used to determine the outputs on the system).

But I like it; it is not only a nice way of reconciling PCT and
IT, it also supports what Martin seems to have thought all
along, namely, that IT, to the extent that it is relevant to
PCT at all, is mainly relevant to reorganization. So there IS infor-
mation about the disturbance in perception -- but it is only partial
information and it is not information that is used to guide the output
of the control system. The information in controlled perceptions is
about the residual, uncontrolled effects of the disturbance. This
information can be used by the reorganization system (NOT BY THE
CONTROL SYSTEM); it is used to "build" and, then, continually
"tune up" the output function.

Beautiful. So we all win. Martin et al are right because there IS
information about the disturbance in controlled perceptions.
We are right because this information is not used by the control
system itself.

Nice journey.

Best

Rick