Galapagos Iuridae scorpion outlier

pierre deleporte pierre.deleporte at UNIV-RENNES1.FR
Tue Apr 23 17:39:05 CDT 2002

A 10:53 22/04/2002 -0400, John Grehan wrote:

>I was not being asked a general methodological question, I was being asked
>about my decision regarding the Iuridae and the issue raised was over my
>application of  panbiogeographic methodology, not biogeographic methodology
>in general.

I understand that. I still consider that discussing the application of
panbiogeographic methodology requires explicitations and justifications on
the basis of general principles. The discussion of the slightest detail of
application of any possible method requires that.

>As pointed out above, and in other postings, there was no relevant evidence
>discarded. The Mediterranean record was not necessary to describe the
>eastern Pacific track.

So it seems that you were focusing of the detailed "local" description of
an otherwise accepted general pattern. The requirement is thus to have
included all the available relevant evidence for infering that pattern in
the first place. You can of course "describe" some details afterwards, if
it's just commenting the results, not inferring the pattern in the first place.

>So far published objections to panbiogeography have been largely of
>a rhetorical nature. People object to it because it conflicts with some
>other view they hold about systematics, historical geology, fossils, age of
>taxa, molecular clock theories, theories about dispersal, theories about
>method, theories about biological differentiation etc.). What will be
>interesting (for me at least) will be to see the critics actually carry out
>a comparative analysis using panbiogeographic and other techniques to argue
>the case and demonstrate, for example, by that application how
>panbiogeography is deficient in its predictive power, or how its past
>predictions were illusions (a claim already made by one dispersalist), or
>the current spatial correlations (between tracks, with geology) do not
>really exist etc. The Galapagos could be an interesting 'test' case.

"Theories about methods" are relevant, not rethorical in a pejorative sense.

"Comparing methods" is of course recommended, generally.
Just a comment, on quite general grounds (sorry). You insist on prediction,
- 1) explanatory power is important, and in biogeographic retrodiction
explanation requires assumptions ("evolutionary models" sensu lato), i.e.
rules for explaning the present state of the evolving system 'taxa and
Earth" by inferring its previous states to some extent (e.g. present
distribution explained by a center of origin, or past dispersal, or
vicariance events, with or without geological changes a.s.o.).
- 2) predictive power in biological terms requires the possibility of
extrapolating the future of the system (not the past), and checking in the
future if "observations" fit the predictions (long-term predictions need a
little bit of patience...).
- 3) "predicting" that our present explanation is right and will remain
correct for science in the future is in my view a quite different kind of
statement: it is a pronostic for the little history of science (e.g.:
methods will not change), and also it implies an assessment that our
sampling is representative (e.g.: we have gathered sufficiently numerous
and informative data to definitely assess the biogeographic history of this
lineage, or of the biota in this part of the world). But I really wonder if
we have theories predicting the future of human knowledge...
- finally, comparing the efficiency of two biogeographic methods for
treating the same data set, as you suggest it, raises some problems:
    in the perspective (1) above, we have no time machine to check the
validity of our retrodictions, thus no empirical test is possible for the
inferences issuing from different methods. The debate is thus
methodological: which method is the more logical for answering this or that
question ? The debate about the best explanation must bear on the grounds
for explanation.
    in perspective (2), we effectively could predict... and wait and see...
the Galapagos in one hundred thousand years...
    in perspective (3), this concerns the future of science, and thus the
fate of methods, once again, and also the largely impredictible fate of new
discoveries (new or revised data).

On the whole I see little possibility to empirically "test the methods"
with biogeographical data properly. Maybe through predictions in other
fields than historical biogeography per se?

Now I can just encourage you to engage in comparing the treatment of the
Galapagos case with different methods. I think you should rapidly encounter
methodological questions of general relevance (what are the questions at
stake, logical coherence of the approaches, justification of underlying
assumptions...), rather than tests of biogeographical predictions properly.


More information about the Taxacom mailing list