John Grehan jgrehan at SCIENCEBUFF.ORG
Mon Jun 21 08:51:04 CDT 2004

I guess one deficiency of a list discussion is keeping track of the
conversation. The posting below and the other one by Agnes Dettai are
indicative of that problem. My issue with cladistic analysis as
practiced through the art of a posteriori use of an outgroup to polarize
characters is that it cannot do any better with the characters than one
has already determined in the first place. Outgroup polarization of a
character that has been identified in such a way as to qualify as an
apomorphy is fine if the character (or character state if you like) is
really apomorphic. My contention is that one does this character by
character, using the outgroup(s) before the data matrix is constructed.
Then it becomes a matter of the analysis being one of resolving
conflicting groups of synapomorphies. I also content that the use of
multistate characters within the in group is fine if one makes a
determination at the outset about the polarity rather than relying on
the manipulations of the analysis to make your mind up for you. If you
do not know enough about the morphology you work with to decide about
multistate character polarity then I would have little confidence in the
qualify of the systematics no matter how sophisticated the algorithm. My
additional contention for molecular characters is that a posteriori
rooting to polarize and make decisions about apomorphic states
represents the application of a cladistic technique to data that is, of
itself, represented in a cladistic format without necessarily being
restricted to apomorphic states.

As for the data matrix by Dick, characters 2 and 3, for example, are
represented as being absent in the outgroup. That's fine if that is
actually the case. My contention is that the analysis of each character
in the first place is what counts. If either character were represented
in such a way as to overlook its presence in the outgroup (or other
outgroups that should be brought into consideration) then they should
not be there in the first place. The algorithm cannot do that for you.
There are systematic studies, for example, that try to work out the
relationships of the great apes by using gibbons as the sole outgroup,
so claims are made about apomorphic characters that are not sustainable
if monkeys are included.

I will be away for a few days after today so I may not respond to any
postings until later in the week.

John Grehan
John R. Grehan
Director of Science and Collections
Buffalo Museum of Science
1020 Humboldt Parkway
Buffalo, NY 14211-1193
email: jgrehan at
Phone: (716) 896-5200 ext 372

More information about the Taxacom mailing list