Re: testability

[George Engelmann <engelman@cwis.unomaha.edu>]

I tend to agree that the hallmark of science, or more generally objective
knowledge, is testability. Testability is vulnerability to "disproof" or
logical contradiction by experiential evidence. Popper developed his ideas
with the physical sciences and law-like hypotheses in mind and it takes some
doing to apply them to historical sciences like paleontology, but I think it
is still appropriate. Much of paleontology (and virtually any other science
that deals with events remote in space or time, i.e. geology, cosmology, etc.)
is involved with reconstructing past events or phenomena. Although it might be
argued that because direct observation of these events is impossible,
statements about them are untestable, I think this is too narrow a view. Events
have consequences and some of these consequences can still be observed today.
As long as a statement about some phenomenon that is remote in time and/or
space is logically connected by arguments or statements (either stated or
implied) that are themselves testable, to present-day experience and 
observation, I think it can be regarded as testable in the same sense that non-
historical statements are. Popper criticized as unscientific, statements that 
are tautological and therefore irrefutable and hypotheses that are so well 
defended from refutation by qualification that any conceivable observation can
be seen as consistent with the hypothesis. Of course it is possible to 
construct such hypotheses in paleontology, but I think it is far more common
(as in dealing with the behavior of fossil organisms) to offer conjectures for
which appropriate present-day observations are sparse or absent, and/or for
which the connecting hypotheses are poorly constrained, do not make the 
necessary logical connection or are just plain wrong. One might consider a 
speculative statement that has such failings to be frivolous and indeed, I 
think that it is the responsibility of the investigator to make an honest 
attempt to uncover such potential problems to the best of her/his ability if it
is to be taken seriously as science. But even if the evidence and connections
are weak, the conjecture may have some value as the source of a better
constrained hypothesis should circumstances change. The important thing is
to distinguish such raw conjecture from better-corroborated hypotheses and not
use such weak hypotheses as the connecting arguments for still more 
speculative statements without making them explicit. Even in science, almost
anything goes in a bull session such as might take place here, and who knows 
where it might lead?

George F. Engelmann
engelman@cwis.unomaha.edu