Fwd: Re: Molecular taxonomy: on way out?

[pierre deleporte]

>A 10:59 19/07/2005 -0400, John Grehan wrote :
>
>>  Why bother putting anything in that is not prima face a derived character?

Why bother expurging the data matrix from non-informative characters when 
the algorithm doesn't take them into account?
And more importantly, why a priori throwing away non-prima face 
apomorphies, like those possibly issuing from reversals, for instance? Do 
you implement a no-reversal evolutionary model? Why? Some fish seem to have 
lost their scales, you know, and some mammals nearly all their hair, and 
some vertebrates a couple of toes... reversal can occur, it seems.
But you can use a 'no reversal' option with your programif you like, this 
would still not be a matter of molecular data against morphological ones.

>>I do not need a mindless algorithm to include
>>and polarize the character. It is that simple.

The algorithm does that anyway, it's that simple and much faster.

Now, if you implement a clique analysis with no reversals, I bet that it 
can occasionally be 'that simple' by hand... the question is now to justify 
your model of character evolution:
'Character, Thou shalt not reverse!'

>>I take the view that these are parallelisms. I know one can play
>>with this in algorithms, but for my analysis I would treat the character
>>state as absent in the outgroup.

Strange indeed. Thick molar enamel is not absent in the outgroups, it is 
polymorphic (present or absent according to outgroup species).
Facts are block-headed.

>>I agree that individual characters can not be phenetic,

Great!???

>>but one can have
>>a group of characters that provide a measure of overall similarity

No, it is the algorithm that provides measures of similarity in characters 
between taxa, only if the algorithm is phenetic.

>>and dumping a cladistic algorithm on top of them does not change the dirt
>>below.

...not great finally: seems you still don't get the point at all. 
Individual characters cannot be phenetic, but a bunch of individual 
characters cannot be phenetic either. Characters provide no measure of 
similarity by themselves. Algorithms do.

>>As for nelson, perhaps he would not agree with all of my views
>>on cladistics, but then again he apparently does agree with my critique
>>of molecular techniques being pseudo-cladistic.

Some molecular approaches are, but not all of them, some are even strictly 
cladistic in the traditional way, and you don't seem able to make the 
difference. Apparently because you persist arguing data matrix instead of 
algorithm.

>>  Everyone is entitled to their choice.

This sounds like a classic 'post-modernist' way to reject science and 
rationality. But logics is not a matter of choice.

Phenetic analysis is not a question of how you fill your data matrix, and 
cladistic algorithms do perform cladistic analysis with morphological and 
molecular data as well. Hence you have no choice, unless you reject 
rationality, which is not recommended.

Unless you charge words with quite unusual meanings, which should be 
explicited in order to avoid a dialogue of the deaf.

Or prove me wrong (see also Richard Jensen for comments on computing).

Pierre


Pierre Deleporte
CNRS UMR 6552 - Station Biologique de Paimpont
F-35380 Paimpont   FRANCE
Téléphone : 02 99 61 81 63
Télécopie : 02 99 61 81 88