KR control, emotion

[From Rick Marken (961014.1100)]

John Anderson (961014.1035 EDT) --

I think that the r3 in your diagram is unnecessary. And your e3 seems to
function in exactly the same way as my r2; it changes the "bias" for
classifying the perception (p2). My r3 was setting a reference for an
"intrinsic" error (the average error in the KR control system) and adjusting
the perceptual function (S2) on the basis of the size of that error.

Your model doesn't adjust the perceptual function; only the "bias" (e3 in
your diagram) for classifying the perceptual variable (p2) as one thing or
another. Your model adjusts this "bias" based on error in the KR control
system. This is probably how subjects actually do behave once they have been
"trained up" in a typical signal detection task. There is usually no way
for subjects to improve their performance (get the "correct" response more
often) by improving their perceptual functions. In the typical detection
task, for example, physical measures of the "signal" and "noise" events are
made to overlap so that the subject _cannot possibly_ produce perfect
performance (control KR perfectly). So all the subject can do, ultimately,
is change the "bias" for saying that the stimulus is one thing or another.
This will help with the control of KR if most of the stimuli are of one type
rather than another. Many signal detection studies are aimed at measuring
a change in bias as a function of the relative probablity of the different
types of stimuli.

Re: The emotion(al) debate

Coincidentlly, there is a front page story today (10/14) in the LA Times
about "new advances" in our understanding of emotions (you can find the
story at http://www.latimes.com/ HOME/ NEWS/FRONT/t000089193.html). This is
part of series on advances in our underdtanding of the human brain. It's
interesting that, among all the new advances, there is no mention of the most
important advance of all: our understanding that the brain is an input
control system. You might think that such an advance would influence the
theories people develop to explain emotion. But, the "input control"
adavance seems to have had no influence on theories of emmotion at all. In
fact, we get a fellow named Tooby (from my PhD alma mater, UC Santa Barbara)
describing emotion as "a mode of operation for the brain, helping it call up
appropriate behavioral programs". Now I'm sure Bruce Abbott and Hans Blom
will point out that Tooby's "behavioral programs" are really perceptual
control systems (after all, a famous psychologist like Tooby wouldn't make
the ridiculously elementary mistake of thinking that the brain programs
behavioral output) but I was surprised at how little theorizing there was
(actually, there was _none_) about the relationship between emotional
perceptions and the control of perceptual inputs. Nevertheless, the people
making the "new advances" in our understanding of emotion have discovered
that there is a "... constant exquisite shifting interplay between the
physical substrates of our brain and the thoughts and emotions which they
emit". Yep. Ain't that just like a control system: emitting thoughts and
emotions all over the place.

Seriously, though, I find it very hard to believe that it is possible to make
even a small "advance" in our understanding of _anything_ about human nature
(such as human emotions) without an understanding of humans as input control
systems. If one's premise is wrong (and the premise of all work on emotion is
that the brain is an output generation system) then how can one's conclusions
be right?

Best

Rick