Re: [Phys-l] climate vs weather; was "why and how"

[John Denker <jsd@av8n.com>]

On 12/22/2010 12:02 PM, marx@phy.ilstu.edu wrote:
> I spend a lot of time in my courses explaining the difference between
> models and theories.  To me, models are competing theoretical
> frameworks to explain a given phenomena.  The competition is played
> out in laboratories that test model predictions and in the offices of
> theoreticians that work to remove inconsistencies.  Like the
> Highlander, in the end, there can be only one.  That successful, 
> remaining model is the theory.  I explain that a theory is as close
> to truth as we get in science, but it is still subject to further
> testing and may be altered or thrown away if new information comes to
> light.

1) You are, as always, free to define your terms in whatever way 
 suits you.

2) Still, it pays to beware of the following scenario:  Suppose you
 define "theory" and explain your meaning ever-so-carefully, and the
 students learn it ever-so-carefully.  Then ... they leave class and 
 are surrounded by other people who define it in another way and use
 it in a third way, without being nearly so careful to explain what 
 they mean.  This means the students are at risk of misunderstanding 
 other people, and being misunderstood by other people.  On the other 
 side of the same coin, there is the risk that your definition will 
 not "stick" after the students leave the classroom.

For slightly different reasons, a highly refined definition of "model"
is almost certainly not worth the trouble.  There are many words that
scientists use to describe models and such.  There are differences 
in the connotations, but the differences are trivial compared to the 
commonalities.
    * Gauss’s law.
    * Newton’s method.
    * Simpson’s rule.
    * Euler’s identity.
    * Parseval’s theorem.
    * Maxwell’s equations.
    * Pascal’s principle.
    * The Euclidean algorithm.
    * The Tolman relation.
    * The Kutta condition.
    * The Cauchy criterion.
    * The quadratic formula.
    * The inflationary model.

The important goal is to understand how science is done, and more
generally to understand how critical thinking is done.  A hairsplitting
exegesis of these words is not at all helpful in advancing this goal.

=======================

This is a tricky business, and I don't claim to have all the answers.  
 *) There are some words that are so important that it seems IMHO 
  worth the trouble of defining them "just so" and defending the
  technical meaning against all the pressures that would tend to 
  distort it.  I put energy and entropy in this category.  
 *) On the other hand there are some words that are no longer very
  important, and for which the technical meaning is already quite
  ambiguous and/or misleading. In these cases I am content to give
  up and simply avoid using the words.  I put "heat", "adiabatic", 
  and "theory" in this category.  They are not worth the trouble.
  There are other ways of saying what needs to be said.
 *) And then there innumerable words that have multiple meanings,
  in a way that people are content to live with.

> Some have manipulated data and 
> outright lied to protect their public story. 

I have looked into this.  This story has been wildly exaggerated.
When Palin accuses the scientists of manipulating and distorting
the facts, it just cracks me up.  It reminds me of when the Bush
campaign was horrified by Kerry's wartime conduct.  It reminds 
me of when Shirley Sherrod was vilified for being a racist.  It
reminds me of "The Rabbits Who Caused All the Trouble" by James
Thurber.

>  Like the
> Highlander, in the end, there can be only one.

That's not true at all.  It is exceedingly common to have multiple
models aka theories that coexist.
 -- m g h is an entirely appropriate model in many cases
 -- G M m / r is an entirely appropriate model in many cases
 -- 8 π Tμν is an entirely appropriate model in many cases

There is sort of a hierarchy in the previous example, but this is
not necessary or even typical.  
 -- Sometimes a liquid can be modeled as a dilute solution of 
  water in alcohol, 
 -- Sometimes a liquid can be modeled as a dilute solution of
  alcohol in water.
 -- Sometimes the solution is not dilute.  The situation is
  inherently complicated, and the simplifications that are
  appropriate for dilute solutions do not apply.

All models are in some ways apt and in some ways not apt.
Deciding which model(s) to use in this-or-that situation
requires judgment and skill.