Re: Dinofest Report #2 (and final)

[Jonathon Woolf <jwoolf@erinet.com>]

Jonathan R. Wagner wrote:

<snip>

>         1) In some cases, the "purpose" of a taxon can be pretty hard to
> figure out. My interpretation (and this is *only* my interpretation) is that
> Arctometatarsalia was named in order to recognize that certain coelurosaurs
> (tyrannosaurs, bullatosaurs, etc.) form a clade distinct from other
> coelurosaurs, rather than being dispersed variously across the family tree.
> Now, as all three taxa share an arctometatarsus, that character is a
> synapomorphy supporting the clade, and the phylogenetic significance of that
> character was a big point in the paper which outlined the hypothetical
> relationship, the taxon was named for the character.

OK, that matches my previous understanding.

>         However, phylogenetic taxa are independant of proposed phylogeny (at
> least theoretically), and the "taxon concept" the namer had in mind is
> irrelevant to their validity. If you think out the definition of
> Arctometatarsalia ("_Ornithomimius_ and all taxa sharing a more recent
> common ancestor with _Ornithomimus than with modern birds") you will see
> that this taxon will always exist, regardless of anyone's hypothesized
> phylogenetic structure for theropods, or when exactly they think certain
> characters evolved.

Then the definition of the taxon is tautological and conveys no useful 
information.

>         2) That said, there is still room to wonder, because, as I noted
> earlier, the original definition was more or less character-based ("the
> first animal to evolve an arctometatarsus and all of its descendants"). As
> such, the character was important to the original definition. However, it
> has been noted that this is not a very stable means of defining a taxon. For
> example, if two clades evolved an arctometatarsus independantly
> ("convergently" or whathaveyou), how would you determine which one evolved
> the character first?

Very good.  That's my exact question.  Unfortunately, you haven't given a good
answer yet.  No matter what alternatives, evasions, exceptions, riders, and 
other
1040-grade complexities you introduce, you cannot get away from the fact that
ultimately, organisms are classified based on physical characteristics, and only
physical characteristics can tell us if a given organism belongs in a given
taxonomic group.  If the feature(s) used to classify the organism are not
apomorphic, then the classification will be flawed.

>         So, Holtz (1996), in accord with the revisor principle (which
> certainly has a place in phylogenetic taxonomy, as he demonstrated)
> redefined the taxon as _Ornithomimus_ and all taxa more closely related to
> _Ornithomimus_ than to modern birds.

Question: how does one tell if a given organism meets this definition?  If you
answered, "by analyzing its characteristics," go to the head of the class.  All
useful taxonomic definitions are ultimately character-based.

> >2. This feature is apparently not synapomorphic among the taxa included in
> >Arctometatarsalia.
>         3) Synapomorphy does not require that the character evolved only
> once. A synapomorphy is a shared derived character. A character which
> evolved in the common ancestor of a certain clade, and seperately in the
> common ancestor of another clade may be synapomorphic for each of those
> clades independantly but not for both together.

Semantic hairsplitting.  I think everyone who read my post knows what I meant, 
and
also knows you're just trying to cloud the issue.

>         5) That the character evolved more than once does not imply that
> some or all of the taxa formerly included in the group don't still
> constitute a clade exclusive of other taxa.

Irrelevant to my original point: if the characteristic for which the taxon was 
named
is not a synapomorphy for ALL species within the taxon, the name is inconsistent
with the facts.

> >3. Even though the relationship after which the clade was named does not 
> >exist,
> >the clade is still valid.
>         10) What is then "the relationship after which the clade was named"?
> As noted below, a clade cannot be named based on a set of taxa, because
> those taxa and those taxa alone almost certainly do not constitute a clade
> exclusive of all other life. Constraining it to being a clade exclusive of
> other taxa in the study is fine for discussion, but useless for taxonomy as
> other taxa will come along later and you must decide what to do with them.

More semantic evasions.

>         12) I believe you are confusing clades and phylogenetic taxa. P.
> taxa are clades, but all clades are not P. taxa. A phylogenetic taxon is the
> recognition of a clade with a name. The definition of a phylogenetic taxon
> is the formula by which you describe the clade by determining which taxa are
> members. A clade, when *properly* recognized as an ancestor and all of its
> descendants is a real entity which exists in nature and can never be
> "invalid". Therefore, a P. taxon can never be "invalid" (unless we goof up
> the process of defining it somehow.

Technically, no.  But if the definition is goofed to the extent that the clade 
has
no useful meaning, both name and clade should be dropped from the literature and
declared passe.

> >Oh, wait, I forgot.  It's cladistics.  It doesn't have to make sense.
>         Perhaps it would be more productive for you to seek resolution of
> your confusion, rather than resorrting to this sort of arrogant posturing.
> Lack of understanding is not a crime, and is nothing to be ashamed of.
> However, flaunting your ignorance in a flourish of self-righteous melodrama
> is downright rude. It says to the world "hey, I'm not ready to be taken
> seriously by anybody."

Nah, if I'd wanted to give that impression I'd have said "it's cladistics, it's 
not
supposed to make sense."  At the moment, as far as I can tell what I said is
entirely correct: you and other hardcore cladists don't care if what you do 
makes
any rational sense.  Claiming that characteristics don't matter when defining 
taxa
is prima facie irrational -- with most organisms and all fossil organisms,
characteristics are the only things we can use for definition of taxa.  You can 
use
any doubletalk you like to evade that fact -- "crown-based clades," "stem-based
clades," or whatever -- but it all comes back to characteristics.  Those are 
facts.
Everything else is interpretation, and no two people will ever interpret it the 
same
way.

Sometimes I wish dinosaur paleontologists were more like 'amphibian' 
paleontologists
-- they at least have the honesty to admit they haven't a clue about how modern
amniotes and modern amphibians are related to each other or to the basal 
tetrapod
groups like the labyrinthodonts.

-- JSW