Slouching towards Bethleham

[From Rick Marken (970920.0850)]

Me:

So why not just explain my mistake; how would a different way of
analyzing the data show that control is model based?

Hans Blom (970918) --

Do your own experiment and see. Where you see only random behavior,
you remind me of that ancient saying "They have eyes but do not
see".

Your passionate attempts to reconceptualize the results of my
"Open Loop Control" experiment
(http://home.earthlink.net/~rmarken/ControlDemo/OpenLoop.html) remind me
of the chilling lines from
Yeats' "Second Coming":

        The best lack all convinction, while the worst
        Are full of passionate intensity.

I never said that behavior in the open-loop phase of the "Open-
loop" control experiment is "random". You imagined this, perhaps
because you think your point (that there can be control without
control of perception) is made if behavior in the open loop case
is _not random_. But "randomness" has nothing to do with it. The
point of the "Open loop" control demo is that control (measured
as RMS error) is nearly perfect in the closed loop case (when
a perception is controlled) and absent in the open loop case.
The results of this demo show that there is no evidence that
people can control _at all_ when they cannot perceive the
variable to be controlled (or a correlate thereof, as in the
"Proportional Control" condition). That is, there is no evidence
of model-based control; it's all control of perception. If you
continue (as I'm sure you will) to refuse to understand this
extremely simple point then please forgive me if I don't show
up with any gifts when you slouch towards Bethleham to be born.

With conviction

Rick

ยทยทยท

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

[Hans Blom, 970923c]

(Rick Marken (970920.0850))

The point of the "Open loop" control demo is that control (measured
as RMS error) is nearly perfect in the closed loop case (when a
perception is controlled) and absent in the open loop case.

Yes, if you define control as correcting the difference between a
perception and a reference. If there is no perception, there is no
control. That is simple logic and does not require a demo.

The funny thing is, however, that if the visual perception is lacking
(in the open loop case), the "generated output" is quite similar to
the output that would have existed if there had been a perception and
if there had been control of it. Whatever you want to call it, there
is something here that needs explanation.

The results of this demo show that there is no evidence that people
can control _at all_ when they cannot perceive the variable to be
controlled (or a correlate thereof, as in the "Proportional Control"
condition).

Given your definition of control, you're entirely correct ;-).

That is, there is no evidence of model-based control; it's all
control of perception.

Yet, as long as there is no perception, there can be no control of
it. If you insist otherwise, I will "continue (as I'm sure you will)
to refuse to understand this extremely simple point".

Greetings,

Hans