[Taxacom] Felsenstein lecture

[Richard Zander]

Felsenstein's lecture was in two parts.

 

First, he expanded on (based on a 2005 paper) a nice explanation for
rapid switching of selectively covarying expressed traits. It was mostly
presented at the microevolution level, but I figure it might join other
explanations for what is apparent evolutionary polyphyly, particularly
the case where a real good taxon appears to be derived twice, once each
from two different ancestral taxa, though each of the different
ancestral taxa has a shared ancestor. Threshold switching of selectively
linked traits is a nice idea, which he attributed to Wright and others
but elaborated himself.

 

Second, he poo-pooed the need for classifications with a diagram that
showed molecular data plus a model yields a tree, which when combined
with phenotypic data yields biological conclusions. This, he indicated,
logically bypassed traditional classification, and the biological
conclusions are all we need.

 

My response:

Award-winner Felsenstein is a statistician. The president of the Soc.
for Systematic Biology, John Huelsenbeck, is a statistician. I've had
short conversations with both, and both are fine people without evil
intent, and may even have an interest in nature. WE put them where they
are. 

 

In the lecture, Felsenstein sloughed off 250 years of taxonomic
enterprise as "phenotypic data." Really? Consider Bentham and Hooker's
Genera Plantarum, and, um, Zim's Golden Guide to Bird Identification
(I'm a botanist). These are just accumulations of phenotypic data? If
not, consider what they then are. I believe that "traditional"
classifications are observations on evolution, using hard-won identified
conservative traits and gaps for describing clusters of individuals as
taxa, and naming these evolutionarily stable groups of individuals at
various levels of organization. My own formula is traditional
classification plus molecular trees yield even more information about
evolution than is in the traditional classification. A kind of
reciprocal illumination can identify problem areas, such as diploids
being derived from tetraploids (paper by Tom Lammers, for which thanks).

 

Felsenstein told me a while back that he had enormous trouble getting
papers published back in the early 80's when he was trying to introduce
statistical evaluation of uncertainty (the bootstrap) in the face of
pattern cladists' objections that the most parsimonious cladogram was
supported fully by Popperian minimum falsification arguments, and needs
nothing else. He had to publish certain, now classic, papers in British
journals because of negative American reviews. Well, here he is in 2008
bruiting from a bullying pulpit a nihilist nonsense that echoes early
cladists' simplistic and monomethodological ideas.

 

*****************************
Richard H. Zander 
Voice: 314-577-0276
Missouri Botanical Garden
PO Box 299
St. Louis, MO 63166-0299 USA
richard.zander at mobot.org
Web sites: http://www.mobot.org/plantscience/resbot/
and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
*****************************