Re: Greg Paul is right (again) - observational analysis and phylogenetics

[Jason Brougham <jaseb@amnh.org>]

>  The cladistics sort of went along, but just good 
> observational analysis was sufficient (as it was to show humans are apes, 
> mammals are 
> derived therapsids, birds and derived dinosaurs, etc before cladistics). 
> Cladisitic analyses can be useful but are inherently limited in their 
> effectiveness and can be exercises in futility that are taken much too 
> seriously, it's 
> really the fossils that count. 
> 
> GSPaul
> </HTML>

Observational analysis may be sufficient to demonstrate that humans are apes, 
but not to produce a phylogeny of the Hominoidea. 

Schwartz and Grehan ( Evolution of the second orangutan: phylogeny and 
biogeography of hominid origins. Journal of Biogeography (J. Biogeogr.) (2009) 
36, 1823–1844)
actually dissent from the consensus that humans are most closely related to 
chimpanzees. They take issue with some methods of molecular analyses, such as 
the pervasive use of molecular plesiomorphies and assumptions used to align 
nucleotide sequences of different lengths. They started with a fairly long list 
 list of features shared between orangs and humans to the exclusion of African 
apes - including tooth enamel morphology, scapula morphology, hair morphology, 
and the foramen lacerum in the palate. They then did a parsimony analysis that 
excluded features found in outgroups, and they found that orangutans and humans 
are more closely related.

Now, I don't necessarily subscribe to their conclusions, and they came up with 
this result through quantitative analysis, but I bring this up as an example 
that a phylogeny is never obvious. The data must be quantitatively analyzed, 
and then robust scientific debate is important for improving the quality of the 
methods and results. New fossils just provide more data, they don't produce the 
phylogenies on their own. If workers in Paleontology use methods that are 
qualitative or subjective, that are not reproduceable, that are not explicit, 
or that can't be repeated by others with improvements and modifications, then 
their contributions are inherently limited.

In my mind cladistics is an effort to develop a system that is as explicit, 
quantitative, rigorous, revisable, and repeatable as possible. In my opinion 
the criticisms of cladistics - such as the assertion that cladistics is 
vulnerable to errors caused by homoplasy - goes double for other methodologies, 
which usually resort to the best judgment of the worker, rather than anything 
repeatable or revisable by better methods. Rigorous methods rescue us from the 
days when one paleontologist says that Archaeopteryx's femoral head is rotated 
anteriorly, and another expert says it isn't, and then that's that.

Jason Brougham
Senior Principal Preparator
American Museum of Natural History
jaseb@amnh.org
(212) 496 3544