[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Cladists versus non-cladists

At 09:43 AM 5/7/97 +0800, Graeme Worth wrote:
>The problem for someone like myself is - how to judge these comments?
>Cladistic analysis seems to have been adopted by most professionals. 
>From my own experience in science, new ideas are not generally 
>accepted just because they are new - in fact the opposite often
>applies! ...

I would hardly call cladistic new anymore.

> there was a time when cladistics was
>unknown and everyone used variations on phylogenetic classification 
>(this is probably the wrong terminology, and someone is bound to
>point this out to me, but you know what I mean). Now, unless
>paleontologists behave very differently from other branches of 
>science, cladistics would not have reached its current position of 
>prominence unless it had a LOT going for it.

Oh, it does.  Its method of phylogenetic inference is by far the best
currently available.  And its approach to classification has the advantage
that the classification can be virtually read off of the phylogenetic tree
(the cladogram).  This, in a certain sense, reduces the subjectivity of

>It cannot be dismissed out of hand, it would seem, without 
>simultaneously dismissing paleontology as a legitimate branch of 
>science. Is anyone seriously proposing this? 

No.  What I am proposing is that, while cladistic phylogenies are of great
value, cladistic classifications lose something of value as the cost of
increased objectivity and standardization.

The more I think about it, the more I realize that general evolutionary
classifications and cladistic classifications serve *entirely* *different*
purposes.  Much of the "argument" comes from the use of the same
nomenclature by both types of classification, making them seem to be
alternative ways of doing the same thing.

This is not, in fact, the case.  Phylogenetic classification (that is the
"cladistic" sort) serves to provide a terminology for discussing
phylogenies, and also a simple textual way to represent phylogenies.  Taxa
based on cladistic principles proved a way to uniquely and unambiguously
speak of relationships and lineages.

Evolutionary taxonomy, on the other hand, provides a way to summarize and
index general information about organisms.  It seeks to optimize both
overall information content and convenience of access (this last is partly
accomplished by keeping the number of names and categories to be remembered

>If not, then it becomes difficult to accept cladistics as the
>accepted method but then cast aspersions on the same method when
>its conclusions do not agree with your own.

It is important to remember, there are two components to taxonomic
practice: producing phylogenies, and deriving classifications.

It is possible to use cladistic methods for step one, and use other
approaches for step two.  [Now, I agree that George is playing a somewhat
different game here - but I reserve the right to disagree with him at any
point :-)].

(Actually, I believe that cladistic phylogenetic methods need to pay more
attention to statistical significance of results, rather than just going
for the set of most parsimonious cladograms, but that is a subtle point,
and doesn't really invalidate the basic methods).

P.S. A recent article in _Science_ described a way to generate likelihood
ratios for DNA phylogenies, and thus a way to produce MLE cladograms
instead of maximum parsimony ones.  All we need now is to find a way to
generalize the method to character-based phylogenies, and a fully
statistically valid cladistic methodology will become available.

May the peace of God be with you.         sarima@ix.netcom.com