🚀 go-pugleaf

[From Rick Marken (990421.2130)]

Me to Isaac:

> Have you figured out the problem with Rushton et al's test
> for the variable controlled when a person walks to a target?

Lazare, Mark (990421)

> Would it or could it be a balance of;
>
> Distance ?
> Effort ?
> Time ?

Ok. I'll explain.

The Rushton et al experiment tests to determine which of two
variables _cause_ locomotion: FoE or target image position. They
test this by seeing what happens to locomotion when subjects wear
prism glasses. Prism glasses cause a change in target image
position but not FoE. So, if FoE is the cause of locomotion then
prisms _will not_ have an effect on locomotion. If target image
position is the cause of locomotion then prisms _will_ have an
effect on locomotion. Rushton et al found that the prisms _do_
have an effect on locomotion so they concluded that target image
position is the cause of locomotion.

All this makes sense in terms of a cause-effect model of behavior.
But it makes no sense at all in terms of a control model of behavior.
The fact that the prisms seem to have an effect on locomotion does
not rule out FoE as a controlled variable. We know from the optics
of the situation that FoE will remain the same (on target) with the
prisms on as long as the subject moves toward the actual target
(as Rushton et al noted) or toward the _image_ of the target (as
Bill Powers noted). So when we see a person "locomote" toward the
image of a  target with prisms on the FoE is the same as it was
with the prisms off.

This was Bill Powers' (990419.1452 MDT) point when he said:

> if a person wearing the prisms maintains the image of the target
> in the center of vision, then the optic flow will be outward from
> that center just as if there were no prisms.

So, for all we know, the subject's _were_ controlling FoE in the
Rushton et al experiment. The FoE is apparently about the same in
both the prism and non prism conditions. But we don't know whether
or not this happens because FoE is controlled; that is we don't
know whether FoE is the same in the prism and non-prism conditions
as a result of an active control process or as an accidental side
effect of controlling some other variable (such as target image
position).

The only way to test whether or not FoE is controlled is by
monitoring FoE _while_ it is being disturbed to see if FoE is
protected from the disturbance. This is what Rushton et al did
_not_ do becuase they didn't know how control systems work. So
they certainly didn't rule out FoE as a possible controlled
variable using the test. And the results of their test to
determine whether target image position is controlled are so
noisy that they are completely inconclusive.

Again, Rushton et al missed doing the test because they assumed
that input "causes" or "controls" output; they would have
done much better if they had known than input is "controlled" by
output; then they could have tested to determine which inputs
_are_ being controlled.

Best

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