Seems Clear to Me

[from Gary Cziko 930407.1806 GMT]

Rick Marken (930406.1800) said:

It is clear to me that people will believe what they
want to believe and there's not much one can do about it (I can already
hear the comments on this one --"well, the perception did have SOME
information about the disturbance; at least for the first 70% of the
time; maybe if we put a lag in or something we could reconstruct d from
p, etc etc). It's hopeless.

I think I see why. The only way to know how the disturbance (d) was
related to the perception (p) is to know what d was so that we could
construct the appropriate function between d and p and use it to know d
when given p. But since we don't know what d is (without the function
relating d to p) and since we can't get the function without knowing d,
there is no (useful) information about d in p (unless, of couse we have the
function which we couldn't have unless we already knew d in which case we
wouldn't need the function).

So it seems to me be that there is no (useful) information in p about d,
nor information in o (output) about d. Maybe that's what you've been
trying to say all along. But if so, why didn't you just say it? Seems
simple enough to me.--Gary

ยทยทยท

------------------------------------------------------------------
Gary Cziko Telephone: 217-333-8527
Educational Psychology FAX: 217-244-7620
University of Illinois E-mail: g-cziko@uiuc.edu
1310 S. Sixth Street Radio: N9MJZ
210 Education Building
Champaign, Illinois 61820-6990
USA
------------------------------------------------------------------

From Tom Bourbon [930407.1530 CDT]

   I just logged back on to CSG-L, after the move from Galveston, and
found what appear to be the last gasps of a discussion of "information
in perception about disturbances." From what I saw in three posts (from
Gary Cziko, Bill Powers, and a quote from Rick Marken), all I can do is
wonder what the other side of the discussion was like. The points they
made were all clear to me. Bill's demonstration, in particular, was
telling, as evidence against any notion that p (or o, for that matter)
includes any readily discernible "information" about d. Even a simple
perceptual control system will eliminate the effects of the NET disturbance
on the controlled variable -- with several disturbances combining in
who knows how many ways, and with error signals working through output
functions whose properties vary, and with potentially variable links
between the output function and the controlled variable, and with
the option of the control system changing its reference signal, the
results of the simulation that Bill reported are definitive.
   I wonder what the other half of the discussion was like. (That is
only a musing, not an invitation to resume the debate.)
   Until later.

[Martin Taylor 930415 2310]
200+ messages to read, so not replying to most. But...

(Tom Bourbon 930407.1530)

I just logged back on to CSG-L, after the move from Galveston, and
found what appear to be the last gasps of a discussion of "information
in perception about disturbances." From what I saw in three posts (from
Gary Cziko, Bill Powers, and a quote from Rick Marken), all I can do is
wonder what the other side of the discussion was like. The points they
made were all clear to me.

I think they were clear to the "other-side" participants in the discussion
as well. As one of them, I can say that nothing in the postings from Rick,
Bill, or (since I departed) Gary introduced any concepts about PCT that
were not already acknowledged and understood. What they did (and continue to)
demonstrate is a lack of understanding of the concepts and constructs of
Information Theory, and an unwillingness to read the words written, in
contrast to the words they imagine to have been written. The discussion
may have been informative to some of the lurkers, but it was frustrating
here, since the original objective was to develop the concepts wel{
beyond the point at which Rick kept arguing against things that were
not as he assumed them to be. Without the initial understanding of
fundamental concepts (such as the non-equivalence between correlation
and information), and the difference between information and reconstruction,
it is very hard to move on to more interesting things. I hope that% some
forward motion in this area will be possible after my return in June.

Bill's demonstration, in particular, was
telling, as evidence against any notion that p (or o, for that matter)
includes any readily discernible "information" about d.

The nub of the problem is in those words "readily discernable." They
should not be equated with "usable."

Bill's demonstration with the multiplier may have come out of his head,
but I did suggest it to him. I also pointed out that the relation
between output and CEV is not even a function, being at the mercy of
changes in the environment. Rick was arguing that the IT view had a
problem if the function was non-linear, and I was trying to tell him that
non-linearity didn't matter a jot, and neither did stability of the
function (within limits of bandwidth describable in IT terms.) I provided
a new relationship, which neither Bill nor Rick cared to comment on
when saying IT can say nothing new about PCT. There's a lot of imagination
in the perceptual realtionships involved in reading.

Sorry for intrusive or altered characters in the text here. Phone line
is relatively good, but not perfect, and I don't have time to edit.

Martin