Keith B Miller wrote:
> I have stated this several times in previous posts -- historical science IS
> predictive. Hypotheses are continually being tested by comparing
> expectations of the hypotheses with future observations. It doesn't matter
> that the events being reconstructed are in the past, only that the specific
> observation or data was unknown to the investigator previous to the
> prediction. This is done all the time. In my own research I am
> continually testing my expectation against new observations. If they prove
> out, my confidence in my hypothesis increases, if they don't that
> confidence is weakened. If expectation are frequently not met, the
> hypothesis is abandoned. That is the way all scientific theorizing works.
> The recent discovery of the walking whales from Pakistan are a great
> example within the field of paleontology.
Understanding "prediction" in a narrow sense, i.e., "speaking about it
before it happens", is too narrow for purposes of theory evaluation. As Keith
points out, speaking about some event or phenomenon which has already happened
but before observational evidence for it has been found should certainly count
in favor of a theory.
But the concept should really be even broader than that. In Lakatos'
approach to scientific theories speaks of "novel facts." These can be things
that were previously unknown (like those considered previously) but also can
include facts which were previously known but which were not adequately
explained by previous theories, & which played no role in the construction of
the theory which the "novel fact" is taken to confirm. A classic example is
the precession of the perihelion of Mercury. It had been known for ~60 years
before Einstein developed general relativity, but counts as a "novel fact"
because Einstein made no use of this phenomenon in the development of his
Nancey Murphy's Theology in the Age of Scientific Reasoning, pp.66-68 discusses
this concept of "novel facts".
George L. Murphy
"The Science-Theology Interface"
This archive was generated by hypermail 2b29 : Sat Dec 08 2001 - 08:01:29 EST