Simonized PCT

[From Rick Marken (960908.1800)]

Bruce Abbott (960907.1455 EST)

What I _am_ suggesting (and have provided the evidence to support) is
that Herbert Simon, writing in 1968, was among those who recognized
that the complex patterns of behavior evidenced by purposive systems
reflect more the complexity of the environments in which they operate
than of the systems themselves.

In scientific work it is considered proper form to acknowledge all
those who expressed a particular idea that one is now repeating;
in a paper I might make the following statement:

   The complexity of the behavior of a purposive system reveals more
   about the complexity of the environment in which the system must
   operate than of the system itself, which indeed may be quite simple
   (Powers, 1973; Simon, 1969).

This would be a VERY misleading attribution because it would suggest
that Simon and Powers came to the same qualititive conclusion for the
same quantitative reason..

Simon's conclusion is based on an input-output model of behavior.
Such a model says that an environmental input, s(t), leads to
behavioral output, o(t), via the organism. The "organism function",
f(),that converts s(t) into o(t) can be quite simple; for example,
f() can be a simple multiplier, k, so that:

(1) o(t) = k s(t).

If k were the actual organism function, then it could be correctly
asserted (as Simon does) that the complexity of behavior, o(t),
reveals more about the complexity of the environment, s(t), than
about the the complexity of the system itself (the simple
function, f()).

PCTers come to the same conclusion as Simon, but for very different
reasons. The PCT conclusion is based on a control model where
environmental inputs are disturbances to controlled variables.
So environmental inputs, s(t) lead to behavioral output, o(t) via
the mutual effect of both variables on the controlled variable.
PCT shows that the organism function, f(), has nothing to do
with the relationship between s(t) and o(t). In a control loop:

(2) o (t) = -1/g(s(t))

where g() is the "environment feedback function" -- the effect of
the output variable on the controlled input variable. The fact
that the relationship between output and input depends on the
nature of the environment and not on the nature of the organism
(as shown in equation 2) is the reason why the control theorist
says that the complexity of behavior reveals more about the
complexity of the environment (in this case, about both s(t) AND
g()) than about the complexity of the system. In fact, PCT shows
that the complexity of behavior (output variations) tells us
_nothing_ about the system; it tells us _only_ about the nature
of the environment in which the system behaves. PCT shows that
the only way to learn something about the complexity of the
system is to learn (using The Test) about the perceptions it is
controlling.

I don't recall Simon ever advocating the abandonment of cause-
effect based psychological research in favor of research aimed
at the discovery of controlled variables. But this is what he
surely would have done if he had come to his "insight" about
behavioral complexity through the doors of PCT.

Bill Powers (Sat, 7 Sep 1996 15:27:57)--

The problem with going up a level of verbal abstraction is that
you could be referring to very different -- even opposite --
underlying ideas... So someone may agree with you that behavior
is purposive and that the complexity of behavior reflects the
complexity of the environment, yet hold a theory of behavior
that is completely antithetical to yours.

Bruce Abbott (960908.1025 EST) --

Simon assumed only that the system has an inner organization
that allows it to perform some function (its purpose) and a
permissive environment in which that function can be carried out.
This is enough to conclude that the behavior of the system
reveals almost nothing about the mechanism through which the
function is carried out, that the variation in its behavior
reflects mainly the characteristics of the environment.

Perhaps my discussion above gives an indication of what Bill
might mean by "very different -- even opposite -- underlying ideas"
The idea underlying Simon's statement is that a simple organism
function (an input-output rule like "multiply inputs by k")
could underly behavior (output variations) that appears complex;
the complexity of behavior, according to Simon, is a result of
the complexity of the stimulus that "drives" output. The idea
underlying the PCT version of the same statement is that the
organism function has nothing to do with the complexity of
behavior; the complexity of behavior, according to PCT, is an
observer's interpretation of the side-effects of the control
of perception.

Simon may or may not hold to an input-output model of behavior;
judging from what I have read he is familiar with closed-loop,
negative feedback control

Many people are familiar with closed-loop, negative feedback control.
You should know from your experience on CSGNet that familairity
with closed-loop, negative feedback control is no guarantee of
familiarity with the fact that purposeful behavior is the control
of perception.

Nevertheless, I think he makes some interesting points that PCTers
would find illuminating, if they weren't so damned frightened of
other people's ideas. Unfortunately, if I were to present these
points, all I would get is a bunch of defenders of the faith
figuratively holding their fingers in their ears and shouting
"la la la la" at the tops of their lungs to drown me out.

Why are we "defernders of the faith" when we point out how what
Simon says is inconsistent with the idea that behavior is the
control of perception?

Since Simon had no idea that behavior is the control of perception,
I find it difficult to believe that there might be anything
"illuminating" about what Simon has said or about what he has
discovered in his research (most of which is based on the ol'
cause-effect model of behavior). But feel free to present some
point Simon makes that you think a PCTer would find "illuminating".
I should think you would find it very amusing to watch the defensive
behavior of a frightened "defender of the faith" (like me)
confronted by such illuminating tidbits. Heck, I find it amusing
to watch you defend your faith. If I didn't, I wouldn't bother to
disturb it so much;-)

Bruce Abbott (960907.1540 EST) --

What Simon (coauthoring with Vera, who was first author, so
we're not even sure whose opinion is being expressed) may have
said in 1992 has no bearing on what he may or may not have
said in 1969. _That's_ life.

So Simon understood control of perception in 1969 but fell back
into the cause-effect model in 1992? I guess Simon's "illuminating
idea" didn't illuminate Simon himself all that much;-)

francisco arocha (96-09-07; 9:34 AM EST) --

My point is that Simon's idea of an organism is not a that
of a control system, but that of an input/output system,
notwithstanding his occasional suggestions, such as that
in his '69 book. Sensory information comes in, there is
some processing in the middle where the incoming information
is matched to some old information in LTM, and information is
sent down for the proper action. This is the idea that I get
from reading not only his '93 paper I cited, but also
his '72 book on problem solving with Newell, his chapter in
the handbook of cog sci and other papers (it is long list).

I think you've got it, Francisco! Another terrific post.

Best

Rick