Positivism in evolutionary science

[Richard Zander]

James Francis Lyons-Weiler wrote:

>         I tend to balk at the positivisitic position that the processes of
>         evolution (be they simple, complex, hierarchical, cyclical,
>         neutral or Darwinian) will have occured in just the right way to
>         allow us to summarize the geneaological relationships among
>         organisms with a parsimony model, for example. Note that I am not
>         questioning the robustness and validity of the evolutionary
>         paradigm -  just what one can correctly expect to do with it.
>
>         This position in not vacuous iconoclasty, but is, I rather
>         hope, a first step (among many first steps) to identify the
>         limitations of how we make historical inferences.  Once
>         the limitations are known, then we can try to deal with
>         them directly.  It is (seems to me) positivistic to
>         assume that we can do our best while ignoring the pitfalls.

I agree with the problems you cite above, and that an analysis of those
limits is sorely needed.

My encyclopedia-in-one-volume says positivism is a philosophy eschewing
explanation and metaphysics in favor of description of phenomena. This
does seem rather extreme, but my own problem with explanation in
phylogenetic inference is that it is used to get us to accept the
most-adequate hypothesis from a group of competing hypotheses instead of
demanding, as we should, that a hypothesis should be adequate just
standing alone. The best (most-adequate) hypothesis does, in fact,
explain a data set as well as the available information allows, but the
best hypothesis, if not having more evidence for it than against it
(P>.5), is not adequate in view of the number of other almost-best
hypotheses. This use of explanation in phylogenetic analysis is
sophistry at best, and like least falsifiability, maximum likelihood,
and maximum posterior probability, is just a ploy to substitute an
attainable goal for a presently unattainable or rarely attainable goal
(a probabilistic hypothesis of evolution).
  A main player in this (aside from Hull) is:
Salmon, W. C. 1971. Statistical Explanation and Statistical Relevance.
Univ. Pittsburgh Press, Pittsburgh, Pennsylvania.
Check out, also, http://dur.ac.uk/~df10www/teaching/exp1.html
and http://dur.ac.uk/~df10www/teaching/exp2.html
  Statistical relevance works in medicine when increased risk (increased
probability of event though less than .5) is associated with great loss,
but its use in phylogenetics is inappropriate.


-
--

*******************************************************
Richard H. Zander, Buffalo Museum of Science
1020 Humboldt Pkwy, Buffalo, NY 14211 USA bryo at commtech.net
*******************************************************