Re: CLADISTICS AND PALEONTOLOGIST(S) OF THE CENTURY

[chris brochu <cbrochu@fmppr.fmnh.org>]

Ladies and Gentlemen,
        I am sympathetic with the concept of reserving this listserver for
dinosaur topics, but as a cladocentrist I feel obligated to reply.
Moreover, there are dinosaur-related issues in this (see below).






>Hm.. Programs?  As in, sets of instructions that run inside a computer?
>Are you
>familiar with the acronym GIGO?


Yes, I am.  Are you familiar with the word "explicitness?"  I think that's
the single greatest strength of modern phylogenetics - matrices are
*published.*  Don't like my results, or Tom's, or anyone else's?  Here's
the matrix, here's the list of specimens examined during its construction,
and here's the specific algorithm I used - have at it.

The tradeoff is between divorcing yourself from "gut feelings" and avoiding
the authoritarianism that characterized earlier phylogenetic methods.  I am
not talking about strong feelings, as much as the fact that relationships
now must be supported by explicit data - not just someone's say-so.

As for the GIGO criticism - I find it a complement that someone can
actually make that accusation about my work (not that it's valid or
anything).  Fifty years ago, one would simply weigh the reputations of
competing authors against each other.  Whatever garbage was in the analysis
could never be localized.  Now, we can actually look inside the data set
itself - surely a better way of doing things.  If there's garbage there,
it's the explicit nature of modern phylogenetics that allows us to find it.




  Are you familiar with the computer
>analyses that
>show a bird-mammal grouping because endothermy is assigned a high
>weighting in
>the analysis?


Actually, there are two classes of results that did this, and none of them
actually placed any kind of weight on endothermy.  There are several
molecular data sets that do this blindly, probably because of some sort of
long-branch-attraction problem; and there are a handful of morphological
analyses, such as those of Gardiner or Lovtrup, where the biggest weakness
was an explicit avoidance of fossils that would overturn their
relationships.  Physiology was not differentially weighted.

THis is relevant to a dinosaur discussion, by the way - it was Gardiner's
point of view - and that of several of his contemporaries - that fossils
could never overturn a phylogeny generated by living taxa, and so he ran
his amniote matrix without fossils, adding them by hand after the fact.
You should see where they ended up - dinosaurs with crocs, and other
fossils in odd places - turtles and dicynodonts, for example.  It didn't
even pass a straight-face test.



>
>> That is completely different from naming clades, which is obviously what you
>> are bungling up with.  Clades are defined on ancestory because they are
>>stable
>> as opposed exclusive membership based.
>
>Which is a pretty useless way of doing it, since you don't _know_ what
>organisms
>are ancestral to what other organisms, and never will.  Call me
>old-fashioned, but
>I think a method of classification should be useful in telling you
>whether a
>specific organism belongs to a specific group or not.


None of the currently-used definitional criteria will tell you that.  It's
the taxonomic hierarchy that provides group membership.  There's a
difference between group membership, group divergence timing, group
diagnosis, and group name - only the first says anything about who belongs
in the group.

As for our lack of known ancestors - this is no different than when
character-based definitions were used.


[shortened]
>
>And maybe we shouldn't keep it.  Likewise the other examples you name --
>in
>ceratopsids, it always struck me as illogical to call something that
>doesn't have
>either horns or head frill a "ceratopsid."  Ceratopsids are _horned_
>dinosaurs --
>and for the sake of simplicity, anything descended from a horned
>dinosaur, like
>_Pachyrhinosaurus_.  Even protoceratopsids could be reasonably called
>horned
>dinosaurs, since the male _Protoceratops_ has an incipient horn on its
>nose.  But
>_Psittacosaurus_ isn't a horned dinosaur and shouldn't be placed in
>Ceratopsia.

Appropriateness has never been a criterion for group membership - even in
the old Linnean system.


chris

-=--=--=--=--=--=--=--=--=--=
Christopher Brochu, Ph.D.

Postdoctoral Research Scientist
Department of Geology
Field Museum of Natural History
Lake Shore Drive at Roosevelt Road
Chicago, IL  60605  USA

phone:  312-922-9410, ext. 469
fax:  312-922-9566

cbrochu@fmppr.fmnh.org