[Hans Blom, 950522b]
a counterexample of "control of perception"
(Bill Powers (950519.1700 MDT))
Hans says control of perception doesn't feel right. What we need
are some demos that show exactly why we say it is perception
rather than the external situation that is controlled. This
means doing something to alter the perception without altering
the external controlled variable, and showing that the external
variable is made to _change_ in just the way needed to keep the
perception from changing.
" .. alter the perception without altering the external controlled
variable ...". Remember my demo? We have done that. My demo includes
a noise term (vt) that corrupts the perception only, NOT what happens
in the world (that is modelled by the term ft). Set vt to a non-zero
value, and see what is controlled: the world (xt), NOT the perception
(y).
And that is exactly what we want: the perception is disturbed and
should not be taken as an exact representation of what is out there.
It is not the (noisy) PERCEPTION that should be controlled.
And whether the perception is corrupted by white noise or a sine wave
or some such is not important. Experiment: replace
y := xt + normal (vt)
in procedure observe by
y := xt + vt * sin (run/5)
and set vt = 0.5 and pvv = 0.25. For a clearest demonstration set ft
= 0.0 and pff = 0.000001. You will find that after a short learning
period xt (the "world" variable) starts to closely track xopt, but
that the observations y (the blue sine wave line) DO NOT track xopt.
So, "control of perception" does not apply to model-based control if
the perception is disturbed and the model adjustment algorithm is
given the information that it is. And that is exactly what we would
want under these conditions.
The question remains whether the perceptions of organisms are noisy
and if so, whether that is taken into account in the processing of
the observations. An example that comes to mind that might be an
example of filtering out sensor noise (even sine wave sensor noise)
are the sudden single frequency beeps that once in a while are heard
in one of the ears and that have a clearly internal origin. The beeps
-- at least in my case -- are canceled fairly rapidly so that "the
world out there" can once again be heard.
Greetings,
Hans