Re: scientific methods

["John S. Denker" <jsd@MONMOUTH.COM>]

Michael Bowen wrote:
> In many cases, community college science courses will be flatly
> rejected for transferability to certain public universities UNLESS the
> course's catalog description explicitly states that the "scientific method"
> (whatever that is) will be taught or illustrated,

Ugh.

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

A good principle to apply in such situations is that it is
important to give the customers what they want, keeping in
mind that what they really want and really need probably
differs greatly from what they initially say they want.

They say they want "the scientific method" but what they
really want is for us to teach people to think clearly.
People aren't born knowing how to do that.  Learning to
think is the most important objective of the physics
course, far more important than any particular facts
and figures.

There _do exist_ things we all recognize as scientific
behavior, taking scientific approaches to problems, etc.,
and we probably pretty much agree what they are, although
it might be hard to come up with a concise-and-complete
list.  By the same token there are certain unscientific
behaviors that we recognize with even greater clarity,
and we can give guidelines for recognizing them.

The problem is that the phrase "the scientific method" has
been hijacked.

> the "scientific method" (whatever that is)

I understand the sentiment -- but the situation is
even worse than that.  We _do_ know more-or-less what
they mean by "the scientific method" but we recoil from
it in horror.

It refers to a certain mode of hypothesis testing that
is not diametrically wrong, but is merely an _approximate_
description of a small _subset_ of the legitimate methods
of science.  To refer to it as "the" scientific method
is appalling.

The version found in the sci.skeptic FAQ is probably
reasonably representative of what is found in textbooks
nowadays:
  http://home.xnet.com/~blatura/skep_1.html

This version is open to many criticisms:
 -- it uses the word "theory" in a way that agrees neither
with the way non-scientists use the word, nor with the way
scientists usually use the word.
 -- it isn't useful because it doesn't explain !!how!! to
formulate a hypothesis.  There are provably infinitely many
possible hypotheses, and you will provably never have enough
data to choose between them, unless some additional processes
are brought to bear.
 -- it doesn't have a realistic view of "discrepancies".  There
are always discrepancies.  It doesn't have a realistic view of
probabilistic predictions.
 -- it misses the concept of competition between _multiple_
hypotheses to find one that is better than the others, or at
least good enough, without needing to pretend that one is "right"
in absolute terms.
 -- et cetera.

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

Another principle to apply is that you can't replace
something with nothing.  If a two-year-old is running
with scissors, don't tell him NO, give him something
else to play with.

Those of us who don't like the usual discussion of "the"
scientific method ought to come up with something better.
There has been lots written on the subject.  Entire books.
A lot of what has been written is nonsense but some of it
is pretty good.
 -- I particularly like Feynman _The Character of Physical
    Law_
 -- There are some important insights in Kuhn _The Structure
    of Scientific Revolutions_

What is needed, and what I haven't seen, is a sensible
one-page discussion of the main points.  This ought to
be doable.  (If it's not doable, somebody should explain
why not.)  Any volunteers?