dispersal fantasy

[John R. Grehan]

Pierre Deleporte wrote:

>There is independent contribution from biology: the genetic datations and
>phylogenies, confronted with the geological scenario, and allowing to check
>for inconsistencies under some interpretative assumptions ("model"). Also,
>everything is always interpreted in the light of current knowledge. Cannot
>be otherwise.

The way the paper seemed to work was that the 'contribution' of biology was
dependant on its degree of conformity to a geological narrative so the
contribution was not independent (.ie. there were no biogeographic models
generated from the phylogeny). Similarly, the genetic information was
another narrative outside biogeography. In reference to current knowledge,
it is possible for a method to generate predictions that go beyond current
knowledge.

>Well-argued "narratives", i.e. historical inferences, may effectively at
>some moment be considered as belonging to background knowledge for further
>explanation. If the narrative is revised, then background knowledge is
>modified and the scientific explanation may change. No methodological
>problem, I think.

If one is confining a research program to making up stories based on
current knowledge rather than advancing knowledge then I would agree that
within that constraint one may say that there is not a problem. If one
prefers a method that is not constrained in this way, then there are
choices available.

>So you require that historical inferences, reconstructions of the past,
>proceed from, or participate in, generating predictions ? Unless they would
>not belong to science ?
>And so, retrodiction (historical biogeography) must generate prediction ?
>Or be thrown away from science ?

I would view a science that is entirely limited to making up explanations
(stories, narratives) in the light of current knowledge to be somewhat less
interesting than a science that advances knowledge. I'll leave it to the
authority of philosophers to decide whether the former is science or not.

>come from historical approaches in themselves. They must come from elsewhere.
>I have no "historical biogeography predictions". Not a matter of taste, but

Perhaps my views are not substantiated philosophically. Whatever the case,
a panbiogeographic approach, for example, provides criteria for
establishing spatial homologies for distributions that allow biogeographic
predictions to be made without having to appeal to historical scenarios
from other disciplines. Such predictions may or may not agree with other
predictions about the past or present. Agreement may be viewed as
corroboration - and perhaps interesting. Disagreement may lead to new lines
of investigation or analysis. Whether this is a better method than
speculating possibilities according to various narratives will be a matter
for personal choice. At present the majority of biogeographic practitioners
appear to choose the former.

John Grehan

Frost Entomological Museum
Pennsylvania State University
Department of Entomology
501 ASI Building
University Park, PA 16802. USA.

Phone: (814) 863-2865
Fax: (814) 865-3048

Frost Museum
http://www.ento.psu.edu/home/Frost/index.html