[From Bill Powers (950910.1820 MDT)]
Bruce Abbott (950908.2000 EST) --
I'm reminded by Dag Forsell that there was also a loose end in your post
that I meant to comment upon. Lifting the quote from his post:
Yep! I'll take a good-ol' slimy, wiggley animal over an
elementary particle any day -- snark over quark, as it were.
What I want to know is, why do physicists call their field one
of the "hard" sciences? It's damned _easy_, I'd say, compared
to unraveling the mysteries of living, behaving organisms.
Even then, they haven't got it right: relativity and quantum
mechanics are fundamentally incompatible theories. And now
what do they give us? _The Emperor's New Mind_ and
fish-particle physics.
I contend that physics has succeeded not because its subject matter is
easy but because its standards for admitting statements into its
structure are set very high. I have asked several times, and ask again:
if physics is easy and the study of living systems is hard, why do
physicists use advanced methods of experimentation and mathematical
analysis when behaving like professional physicists, while psychologists
do not do the same when behaving like professional psychologists?
The problem with physicists turning out things like fish-particle
physics and the various gee-whiz theories of life that we have seen over
the past x decades is that when physicists get outside their own field,
they treat the new subject as an opportunity to take a vacation from the
normal rigor of their science. They start writing like psychologists,
sociologists, philosophers, and so forth. It's as though they feel that
they're dealing with a dumbed-down area of science, so dumbed-down
arguments should suffice to allow them to compete in this less-demanding
arena.
What makes physics and other hard sciences work so well is that they
demand theories that predict accurately, within the limits of our
ability to measure. If there is any systematic contradiction between
theory and observation, a marker is placed into the book of physics with
a note saying "Do not trust what comes after this until this discrepancy
is explained."
In the behavioral sciences, if there is a discrepancy between prediction
and observation the custom is to write it off as a statistical
variation, or to find words with which to describe the observation to
make it seem to agree, retroactively, with the prediction. When
prediction fails, the custom is to point out how complex animal behavior
is, and how difficult it is to control all sources of interference that
might make it seem that the theory is wrong when the theory is really
right. So even if the prediction was wrong, it was probably actually
right.
What do you think would happen in the behavioral sciences if a policy
were set that no hypothesis would be accepted unless predictions based
on it were right within 5% at least 99 percent of the time? That would
be an extremely low standard, completely unacceptable, for the basic
hypotheses of physics, for example the hypothesis that the gravitational
attraction between two particles falls off as the inverse 2.0000...
power of their separation, or the hypothesis that gravitational and
inertial mass are the same.
If your answer is that the effect would be to bring the behavioral
sciences to a halt, I think that answer would be only temporarily
correct. What would really happen, I predict, would be that at least
some behavioral scientists would realize that they have been trying to
fly before they know how to crawl. They would realize that to meet this
standard they would have to do experiments that were very much simpler.
They would have to admit that even concerning the very simplest
phenomena of behavior they are essentially ignorant. Most important,
they would have to give up the pretense of being experts about human
behavior, and become as little children again, studying the simple
things until they were sure they had them right.
Look at Galileo, that great scientist and philosopher. When he wanted to
learn about nature, what did he do? He hiked up his robes, squatted down
in front of a slanted board, and rolled little balls down it, like a
child. All around him, learned men were pontificating (one of them quite
literally) about how nature had to behave, about what the great
principles of nature said, about the vast structure of creation within
which all natural things had to find their place. The idea of getting
down onto a dusty floor and playing with children's toys, just to make
sure of getting it right, would never have occurred to them. Yet it is
Galileo whom we remember and revere, not any of those great men who were
interested in grander things.
If behavioral scientists were to agree that all their predictions had to
be better than 95 percent accurate at least 99 times in 100, they would
simply have to go back to the bases of their science and look for
hypotheses which meet that criterion. They would not find themselves
talking about conditioning or Oedipus complexes or social factors in
mental disease or personality factors versus situations. They would find
themselves playing with rubber bands and asking people to track a moving
finger, or something equally simple. The world would not be very
interested in what they are finding out. They would not be of much use
in helping to solve personal or political or social problems. But they
would be building a science just as powerful as physics.
Then it would become obvious that physics has succeeded not because it
has an easier subject (have you analyzed the motions of a gyroscope
recently?) but because it sets higher standards for what will be
accepted as a fact. When people agree to accept only the very most
probable statements of fact, they begin to accumulate knowledge of a
very high quality, however simple it might be. What they learn can still
be wrong, but the likelihood of its being wrong can be made very small,
so small that the best of the facts will survive incessant testing for
centuries. If, on this solid base, only the very most probable
combinations of these facts are retained and used for further hypothesis
building, the new hypotheses will prove to work far more often than they
do not. And of course those that do not work will also be pruned out by
this implacable standard.
It's simply a matter of natural selection, as Gary "Without Miracles"
Cziko would most likely agree. If the selection criteria are very
stringent, then only the most fit ideas will survive. If they are loose
and permissive, then any monstrosity can make it to the next generation.
The way you get good science is to be fussy about letting bad science
propagate. Where you set the cutoff point determines the quality of what
you get. What could be more obvious than that?
···
-----------------------------------------------------------------------
Best,
Bill P.