Prediction

[X99Lynx at aol.com]

In a message dated 4/17/99 10:08:33 PM, rayhendon at worldnet.att.net wrote:

<<I don't mean this in an argumentative way. I may be just dense on the
matter and fail to see the obvious, but fail to see it ,I do.  The word
"expected" demystifies the statement of predictibility, and puts it in on
more or less a scientific footing,...>>

"Predictability" as it applies to the scientific method is a "term of art."
It's a matter of usage as much as anything else.  Physicists for example will
not only speak in terms of a hypothesis being confirmed by its predicting
results but also judge a theory or "laws" based on "predictive power."

Here are some excerpts from my e-files that illustrate how the term is used:
(these are from abstracts):

"Quantum mechanics does not predict when the individual nucleus will decay,
although if many similar nuclei are surveyed, one can predict what fraction
will decay in any time interval."

"One hypothesis is that the human brain encourages the consumption of foods
that contain valuable calories, while discouraging ingestion of
life-threatening substances.  This hypothesis leads to the prediction that
sweet-tasting foods are edible and nutritious whereas extremely
bitter-tasting materials are toxic, which is generally the case."

"Terzaghi's work introduced scientific methods for predicting the behavior of
soils and developed a basis for rational engineering design."

And here's Victor J. Stenger, a physicist, writing about Newtonian physics:

"The classical paradigm provides us with the means for predicting the motion
of all material systems in the classical domain. Whether continuous or
discrete, these systems of bodies are treated as composed of constituents
that obey Newton's laws of particle motion and the various principles derived
from them. Given the initial position and velocity of the constituent, and
knowing the net force on it, you
predict its future position and velocity."

Other connotations may confuse the use of the word, but when it is used by
scientists, "predictive" will very usually refer to the testability of a
hypothesis or the confirmation of a theory.  As you point out, the evidence
can be in the form of experimental results or observations (Cicero's letters).

"Expectations" seems like a fine word to me, but it is not what scientists
generally use.

Obviously this definition of "prediction" does not have to refer to anything
as grand as where all language will go in the future.  It can be as qualified
as Grimm's Law.  Where there was a *p, we can predict there was an f, under
the right circumstances, more or less.

Historical linguistics is a science that relies heavily upon this kind of
methodological "predictability" both in following the history of a "word" and
in reconstruction.

Regards,
Steve Long