The PCT Revolution

[From Rick Marken (930319.0830)]

Bill Powers (930319.0730) --

I almost fell for Martin Taylor's explanation, Rick. It's only
your persistence that put me back on the track. Martin is a
silver-tongued devil.

Thanks for your support; I'm sure that I must seem like an
uncompromising maniac unwilling to give an inch over a
seemingly trivial point. But to me, understanding that there
is no information about (or a usable representation of) the
disturbance in the perception (or CEV) is the central (revolutionary)
fact of PCT (with respect to the behavioral sciences). It is the
fact that PROVES that the input-output model of behavior -- the
framework within which all work in the behavioral sciences is done--
is WRONG. PCT even shows why it has seemed like the input-output
model works -- disturbance resistance; in the process of controlling
(uninformative) perception, control systems generate outputs that
counteract disturbances to the controlled perception; if you see the
disturbance (like a light shined in the eye) then the output (pupil
size variations) seem to be caused by it. But control systems don't
really work that way -- when organisms are in a closed negative
feedback relationship to their own sensory inputs, then (most of)
these sensory inputs are NOT the cause of outputs; in fact, they
are controlled by outputs, and this control is not accomplished
because of anything informative about the sensory inputs themselves.

Note that those aligned against me on this point are people who
"accept" or at least are sympathetic to the PCT model of behavior
-- Martin, Allan, Oded, Avery. I'm sure there are many others
listening who feel the same way "PCT is great but this Marken
guy is pretty extreme -- imagine, saying that perceptual input
doesn't contain any information about the disturbance. A guy like
that just gives PCT a bad name -- what does he expect me to believe,
that our perceptions don't provide any information about how we
should deal with the world. I like the idea of control of perception
and feedback and all that but I don't see why a PCTer would have
to deny physical reality to try to sell the theory. Having loonies
like Marken around just gives PCT a bad name".

I (Marken) understand how you feel. All I ask is that those who are
interested go over all the discussion and try to understand the
points and the justifications for them. I think you will find
(especially if you do the relevant demos and the relevant math)
that crazy Marken is right. There is no compromise on this issue --
not because I don't want to (I'm all for cooperation) but because I
am correctly describing the operation of a control system;
and if people are organized as control systems (and I think its fairly
safe to say that they are, simply because they are locked into a
negative feedback situation with respect to their own perception)
then what I am saying is true; people compensate perfectly for
disturbances to their own perceptual input, not because there
is information or some usable representation about the disturbance
in the input that allows people to compute, bleed off or whatever,
the appropriate output but because their output is proportional
to the difference between perception and a reference for that
perception -- that's how it works. And the result is that the
perecption is made to match the reference. Ahd since the reference
is ultimately determine by the system (person) themselves, it
means that people determine "what they do" meaning what they
experimence so, in this sense, people are completely autonomous.

Ultimately, I don't think you can fully appreciate the significance
and depth of PCT unless you can understand this apparently trivial
little point:

In a control system, the perceptual signal contains no information
about the environmental disturbances (or system outputs) that are
acting to cause the variance in that perceptual signal. This means
that the perecptual signal cannot be the cause, guide, source of
information, or anything else regarding the appropriate outputs that
should be made by the control system in order to produce the intended
result (the reference value of the perecptual input).

Once you understand this, you can kiss all versions of the input-
output approach to living systems goodby. It might be a painful
divorce but, believe me, it opens the way to a MUCH better life
of science.


Richard S. Marken

"Religion is too important to be left to religions"