cladism's greatest weakness

[Richard Zander]

----- Original Message -----
From: Thomas Pape <thomas.pape at NRM.SE>
To: <TAXACOM at USOBI.ORG>
Sent: Friday, September 17, 1999 3:44 PM
Subject: cladism's greatest weakness


> Reply to Zander:
>
> >Choosing the optimal one against all alternatives is arbitrary.
>
> Taken verbatim it sounds very non-arbitrary to me! We're getting back to
the old >discussion of parsimony.

It will never be an "old discussion" as long as NSF keeps on funding
parsimony studies at the $500,000 to 750,00 level while alpha taxonomy
languishes.

>A parsimonically suboptimal alternative simply remains suboptimal, even if
differences in >optimality to the closest alternatives may be so small as to
be considered insignificant. Yet >how to choose levels of significance  is a
truly arbitrary matter.

You bet. That's a problem, too.

>
> >Bremer support of say 10 on a branch that is 40 steps long leaves 30
steps
> >of data supporting an alternative, and I think reasonably supported
branch.
>
>
> You cannot argue for 'reasonable support' out of context. First, a Bremer
support of 10
>means that ANY alternative not carrying the branch in question will be at
least 10 steps >less parsimonious.

That is also true, but not relevant to fact that there can be a
well-supported contradictory tree given this ratio of Bremer support to
branch length.

>Second, "30 steps of data" is not necessarily 30 steps of support. '

I misphrased that. In any case, it is up to the researcher to demonstrate
that there are no alternative branches with 30 steps of support. I've never
seen Bremer support looked at that closely, but it should be.

>Third, any "reasonably >supported" branch would be discarded in favour of
any such >branch with a (reasobably) higher support.

Why so? Discarding somewhat lesser supported branches is WHY the details of
a optimality-produced cladogram are dubious, QED.

R.

Richard H. Zander, Curator of Botany
Buffalo Museum of Science
1020 Humboldt Pkwy
Buffalo, NY 14211 USA
email: rzander at sciencebuff.org
voice: 716-895-5200 x 351