Theories ... ohhhhhhhh... theories

[From Rick Marken (960731.1300)]

Hans Blom (960731) to Gregory Thomas Wierzbicki:

What to do [about controlling for different theories]? Well, just like you
do now: make us aware of what we're doing. Try to get us to go up a level,
where we can stand "above" our pet theories and can view them more or less
objectively, as just one of the many. Such a colorful world! Why try to
reduce it to drab uniformity?

Spoken like a true liberal theologean;-)

I don't suppose it ever occurred to you that you could subject all these
colorful theories to empirical test? Or is that too brutally scientific?

Best

"Colorless" Rick Marken

[Hans Blom, 960801]

(Rick Marken (960731.1300))

>What to do [about controlling for different theories]? Well, just like you
>do now: make us aware of what we're doing. Try to get us to go up a level,
>where we can stand "above" our pet theories and can view them more or less
>objectively, as just one of the many. Such a colorful world! Why try to
>reduce it to drab uniformity?

Spoken like a true liberal theologean;-)

Wow, Rick, I haven't had such a nice compliment in ages! Thank you so
much! But... what do these words MEAN?

I don't suppose it ever occurred to you that you could subject all these
colorful theories to empirical test? Or is that too brutally scientific?

Dear "Colorless", give me a formula or a set of them and I may be
able to test something empirically. Give me a bunch of words,
unprecise diagrams, or whatever your pet theory may have to offer at
this moment and the test results will be inconclusive, as we so often
saw in the past. You say "My simulations mimick X with 99% accuracy",
I say "But your simulations cannot mimick Y. Try my theory, it
mimicks Y with 95% accuracy". And you say "But Y is not important. We
just want to model X". And I say "But I think that Y is important as
well". "No", "Yes", Ad infinitum. Is that your idea of an empirical
test?

How, Rick, how?

Physics compares theories about ONE very specific phenomenon, where
everyone agrees that X is meant and not Y. And then the theory which
predicts best is chosen, even if it is much more complex than a
theory which is ALMOST as good. Einstein versus Newton. Einstein is
to be preferred although hardly anyone understands general relativity
theory.

If we can agree on some very specific X instead of the fuzzy notion
"human behavior", we may be able to arrive at an empirical test. But
such a test will not "explain" behavior -- only a tiny fraction of
it. I admire your confidence that we will be able to define human
behavior as some X that we can all agree on, but I do not share it.

How, Rick, how?

Greetings,

Hans

[From Rick Marken (960801.1315)]

Me:

Spoken like a true liberal theologean;-)

Hans Blom (960801) --

Wow, Rick, I haven't had such a nice compliment in ages! Thank you so
much! But... what do these words MEAN?

A liberal theologean thinks all theories (religions) are equally correct;
that those who believe in one theory (religion) can profit from the "wisdom"
of the other theories; that value of a theory (religion) is judged by
believers in terms of how well the theory "works" for them.

Liberal theology is the antithesis of science and (therefore) understanding.
A scientist thinks that some theories are far more correct than others; that
there is no "wisdom" to be gained from demonstrably incorrect theories and
that the value of a theory is judged in terms of how well it accounts for
observations, not in terms of how much anyone likes it.

Dear "Colorless", give me a formula or a set of them and I may be able to
test something empirically.

You can start by reading "Mind Readings" and repeating the simple experiments
and running the simple models described in that book.

You say "My simulations mimick X with 99% accuracy"

Yes. My simulations mimic the behavior in various control tasks with 99%
accuracy.

I say "But your simulations cannot mimick Y. Try my theory, it mimicks Y
with 95% accuracy". And you say "But Y is not important. We just want to
model X". And I say "But I think that Y is important as well". "No", "Yes",
Ad infinitum. Is that your idea of an empirical test?

No, Hans. This is my idea of how little you seem to know about doing science.
You have never presented an example of a behavioral phenomenon (Y) that that
can be mimiced by an MCT model and _not_ by a PCT model. You have talked
about "walking in the dark" and you've given vague descriptions of tracking
tasks as examples of tasks that (you say) can be handled by MCT but not PCT.
But you have never shown an experiment where a working MCT model behaves like
a living system and a PCT model doesn't.

If you want to know what I'm after (in terms of experimental evidence
comparing MCT to PCT) check out a paper called "Models and their worlds" by
Bourbon and Powers (it's in an issue of Closed Loop). This paper describes
scientific tests of different models of behavior. The tests are very simple;
what was hard was developing the situations where the different models make
clearly different predictions.

The experiments described in the "Models..." paper show that S-R and
cognitive models of control fail -- period, amen. In what must seem like
egregiously simple experiments, Bourbon and Powers showed that no more time
need be wasted on research based on S-R and cognitive models. I'd like to see
an equally simple set of experiments that let me reject PCT in favor of MCT.
I don't think it should be _that_ hard for you to come up with such
experiments. But you're the one who thinks MCT is a better model than
PCT so you should be the one developing these experiments. But if this proves
to be too much for you, I'll be happy to suggest some experiments myself.

Actually, we have already shown that your MCT model (with the feedback loop
intact) behaves just like the PCT model -- and like a subject in a tracking
task. So your model does just as well as the PCT model (because it is a
PCT model when the loop is closed) in a simple tracking task. Since the MCT
model is an order of magnitude more complex than the PCT model, we're
sticking with the PCT model until observations require the inclusion of an
internal model and a complex adaptive system to maintain it.

Best

Rick

[From Bruce Gregory (960801.1650 EDT)]

(Rick Marken (960801.1315)

to Hans Blom (960801)

A liberal theologean thinks all theories (religions) are equally correct;
that those who believe in one theory (religion) can profit from the "wisdom"
of the other theories; that value of a theory (religion) is judged by
believers in terms of how well the theory "works" for them.

Far be it from me to defend liberal theology, but... Theories
(religions) need not be equally "correct" to contain elements
that we can learn from. All we ever know about a theory is how
well it "works" for us (in predicting the behavior of systems,
living or otherwise).

Liberal theology is the antithesis of science and (therefore) understanding.
A scientist thinks that some theories are far more correct than others; that
there is no "wisdom" to be gained from demonstrably incorrect theories

Many theories have been superceded (Newtonian physics by General
Relativity) but remain useful and productive within the spheres
where they continue to predict accurately. This is totally in
accord with

the value of a theory is judged in terms of how well it accounts for
observations, not in terms of how much anyone likes it.

Since the MCT
model is an order of magnitude more complex than the PCT model, we're
sticking with the PCT model until observations require the inclusion of an
internal model and a complex adaptive system to maintain it.

Can't disagree with that reasoning.

Regards,

Bruce

[From Rick Marken (960801.1430)]

Bruce Gregory (960801.1650 EDT) --

Theories (religions) need not be equally "correct" to contain elements
that we can learn from...Many theories have been superceded (Newtonian
physics by General Relativity) but remain useful and productive within the
spheres where they continue to predict accurately.

I agree completely, of course. So I have to take back my claim that
scientists believe:

that there is no "wisdom" to be gained from demonstrably incorrect theories

I guess there's incorrect and then there's INCORRECT. Newton's theory is now
demonstrably INCORRECT (not the closest approach to the TRUTH), but it is
certainly not incorrect (misleading or useless -- they're using it all the
time over here at satellite central). I guess I was thinking of theories like
S-R, reinforcment and cognitive theories of behavior; these theories are not
only demonstrably INCORRECT; they are also incorrect.

Best

Rick