Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: Modeling reality



On Thursday, Jan 23, 2003 John Mallinckrodt wrote:

> ... Successful new theories in physics must at least make
> predictions that agree in every case where existing theory
> is in agreement with experiment. However, since that
> alone confers no advantage, they must also either be
> simpler or more beautiful than existing theories and/or
> they must make predictions that are at odds with current
> theory and in accord with experiment. ...

Ludwik Kowalski replied:
> This is true for a THEORY. But what about a model?
> A model, I think, is acceptable even if it is not able to
> agree with all known facts.

Whoa, be careful. Remember:
"No theory should fit all the facts,
because some of the facts are wrong."
-- attributed to Fritz Zwicky

J.M's first sentence is fairly careful about
this point, but even so it doesn't go far
enough ... Suppose the old theory agrees
with "facts" A through J, but disagrees with
"facts" K, L, M, and N. Meanwhile the new
theory disagrees with "facts" B through M but
disagrees with "facts" A and N. It violates
the letter (but not the spirit) of J.M's
rule.

There are rules, laws, formulas, equations,
principles, equalities, identities, et cetera.
I don't see much point in drawing fine
distinctions among these concepts.

A model is acceptable if
its predictions agree with facts not explainable by
existing theories and models. It is not as good as a
theory but it is better than nothing. I would say that a
model is a hypothesis, not a theory. Right or wrong?

It depends. Some models are well-nigh exact.
Some models are quite sketchy, qualitative,
and/or hypothetical. There's a continuum.
To my ears, calling soemthing a model doesn't
imply much about where it sits on this continuum.

And as we have discussed before, the word
"theory" is a booby-trap. It has two meanings.
A "mere theory" means speculation; this meaning
has been around for thousands of years. At the
other extreme, scientists use the word "theory"
to mean a grand collection of rules / laws /
formulas / whatever, giving a coherent view of
a subject.

We ought to find or invent a new word for the
latter meaning, to prevent unnecessary confusion
among non-scientists.