[Taxacom] cladistics (was: clique analysis in textbooks)

[John Grehan]

From: taxacom-bounces at mailman.nhm.ku.edu
[mailto:taxacom-bounces at mailman.nhm.ku.edu] On Behalf Of Anthony Gill

I've been very reluctant to weigh in on this thread as I fear that it
will
lead to endless debate with people that have already made their minds
up.
However, I thought I should weigh in on a few points (then run like
hell, so
I don't have to waste more time on it):

> 1) John's methods are not outlandish. In fact I would say that his 
> criteriaare the norm among fish morphologists. 

Never could trust those fish morphologists :)


> He has made a selection of characters
> that he feels he can justify as potential synapomorphies (using
criteria
> defined by Hennig, among others), in so doing, excluding characters
that
> (from his initial experimental design - i.e., with a defined ingroup)
are
> known to be homoplastic. He has not excluded characters that might
produce
> conflicting relationships within his ingroup. His goal is to test
which of
> the characters of choice are congruent and what relationships they
imply.
> Whether optimization (parsimony) is an appropriate method of analysis
is
> another matter. 

This is so nicely put I may have to plagiarize it (I mean use with
acknowledgment)!

> It seems that some in this thread would rather John > sampled ALL 
> characters (whatever that means), include outgroups as ingroups
(again,
> whatever that means), and place his faith in a computer algorithm that
> requires no scholarly appraisal of the characters a priori. As far as
I'm
> concerned, that's a phenetic approach, at least in spirit. I fail to
see 
> how it has anything to do with cladistics.

I guess this is my current position also.

> 2) It may well be professional suicide to engage in 3ia, but that in 
> itself doesn't mean it is an incorrect approach. 3ia is not simply
another 
> method of analysis, but rather a method of viewing the original data.
When
> characters are reduced to the minimal relationships (3 item
statements) 
> they imply, it does not matter whether you analyse the 3 item
statements 
> with compatibility or parsimony analysis ... you will get the same 
> results. What you will not get, however, is an endless shifting of
optimal 
> results (each with a different interpretation of character evidence)
as 
> taxa are added or subtracted from the data sets.

I wonder if Pierre will be telling this person to go and read some
textbooks?

John Grehan