Australopithecus and orangutans

[Herbert R. Jacobson]

Jason, 

First let me say that my statement, "What kind of argument is…," was 
unnecessary on my part. Your argument was clear and reasonable, and accepted 
by many. I should like to blame my response on having lived through the 
taxonomic wars where reasoned and well thought out and argued evolutionary 
narratives were called Just-So Stories, and sarcasm was wielded as often as 
logic. It was hard to be civil. That said I still disagree with your argument 
and your response to my criticism. 


A hypothesis is accepted not because there is a lot of support for it but 
because it has greater explanatory power than any of the competing 
hypotheses. It may be rejected later, not because it has lost support, but 
because another hypothesis with greater explanatory power was found. This was 
why parsimony was selected as the test for competing phylogenies. The most 
parsimonious tree is the tree with least number of homoplasies, or the least 
number of unexplained transformations. And the tree with the least number of 
unexplained transformations that have to be accepted ad hoc is the tree with 
the most explained transformations, hence the tree that explains the most.

Interestingly, the most parsimonious tree is not always the tree with the 
most congruent characters. Character congruency was once proposes as an 
alternative to parsimony, but rejected by most for the reasons given above. 
Here is an example where explanatory power was used instead of support.

At the risk of sounding sarcastic, support for a hypothesis may give a sense 
of comfort or confidence but that's about all it is, a good feeling. Science 
doesn't choose by the majority vote. 

New molecular phylogenies will be accepted because they are better at 
explaining the observations, not because they agree with older work, 
morphological or otherwise.

Further I reject both your statements that seemed to imply I am one that "…
argued …that molecular data contains no historical signal…." and "…chooses 
to ignore molecular data out of principle…." I certainly believe that 
molecular data do contain historical information. I'm just not sure what that 
information is. I hold the position proposed by Gould and others that 
evolution is hierarchical, operating independently at each the species, 
organism, and genetic levels. So I am not certain that uncovering 
evolutionary patterns at the molecular level (historical signals) will reveal 
patterns at the species level, the level which is most interesting to 
morphological cladists. 

I don't think it's stretching it to view molecular evolution of conserved 
sequence as decoupled from the evolution of the organism. After all conserved 
sequences are selected because it is thought they are not influenced by 
selection at the level of the organism. This means there may be more than one 
historical signal. It's analogous to trying to figure out the phylogeny of 
the host from the phylogeny of a parasite when the two can, and often do, 
have different tree topologies. The fact that they may be congruent in some 
cases doesn't help unless you know the answer ahead of time. 

I guess it boils down to whether I am convinced at this point that 
substitution rates in conserved molecular sequences can stand as a proxy for 
speciation rates, and for me the issue is still in doubt.

Herb