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

Re: Classification: A Definition

On Fri, May 18, 2007 at 10:53:56PM +0100, Mike Taylor scripsit:
> Graydon writes:
>  > > No, no, _alpha_ taxonomy, the naming of species and genera.
>  > > Those we will never get rid of.
>  > 
>  > I am uncertain of that.
>  > 
>  > Specifically in paleontology, it's unclear that species level
>  > resolution is possible at all, never mind generally, so the
>  > question ought to be pretty much moot.
> Species or genus -- doesn't matter which, we need one of them.  


We need a label that people can remember and say for a specimen or a
group of specimens.  We don't need to attach a meaning beyond "this
non-empty list of specimens" to the label, though.  I think it could be
argued that in paleontology one *should not* go beyond that, because
there isn't enough information to do it.

In cases there there's a population or genes to look at in terms of
that-which-is-selected-upon there are other species concepts that can
come into play, but even there it helps a whole lot to know *which*
species concept.

So why not something like

Toothyosaurus, HP Fred Derf 1900, AugustInstitution 0001, 0007, 0009a,
Concept 0

(Where "Concept 0" is the "it's dead, Jim" "fossil taxon" species

Things from the current biota can work the same way provided there is
some way to agree on a reference list of species concepts.

> I
> agree that two is overkill (hence my ha-ha-only-serious suggestion of
> abandoning species and retaining genera, which have the advantage of
> being uninomial).  It's either specimen numbers all the way down, or
> you need an arbitrary kind of taxon (i.e. not in general a clade) that
> mean "this type specimen and everything else sufficiently close to it,
> where 'sufficiently' must be interpreted Just Right".

"Sufficiently" is where a lot of the interesting questions are, though,
and that's not something a nomenclature can hope to handle.

I don't see why specimen numbers all the way down is a bad thing; for
living biota, it might start being reference numbers for gene
populations as well as more traditional preserved-critter specimens.

There would be, and would need to be some systematization of, names
defined in relation to those specimen numbers, but that's not very much
different than saying you have to have a holotype.

>  > In biology, it doesn't strike me as genera are any less artificial
>  > than families or orders, or that the "epithet <namer> <date>" style
>  > of species reference wouldn't work at least as well.
> Totally agree.  My fondness for genera is simply that they are more
> PN-friendly than specues due to the uninomial and consequent
> uniqueness.  But as we all know, in the world of dinosaurs, genera and
> species are pretty much the same thing anyway.

I can see people getting upset because for live things there are
applicable species concepts, but I don't think uninomial by specimen
list is really a bad idea even there.

>  > > Above that level I am broadly in agreement that the ranks do more
>  > > harm than good -- which should not particularly suprising given
>  > > that most of what little there is of my publishing record is on
>  > > rank-free phylogenetic nomenclature.
>  > 
>  > I am unclear how the designation of a species doesn't constitute a
>  > phylogenitic hypothesis.  So why are species and genera special?
> Because they are defined on a single anchor specimen.  That is
> pragmatically useful.

They are, but isn't that saying "I hypothesize (or assert) that this
organism is a representative of a population subject to selection which
has currently no named representative"?

Whether or not that is a supportable hypothesis concerning, say
_Tarbosaurus bataar_ or _Nanotyrannus_ has consumed a lot of ink and

[depth slicing a phylogenetic tree for book organization]
>  > Not for an individual book that's trying to document a particular
>  > hypothesis -- some specific tree -- it's not.
> Sure it is.  No-one wants to read an article about the clade uniting
> Lourinhasaurus with Neosauropoda.  Neosauropoda is a fundamentally
> more interesting clade than that one, because its root is the branch
> point into two major groups, rather than one group and a singleton.

"Fundamentally more interesting" is a tricky thing.

If you're after a metric of interest rather than a straight survey of
the tree, though, there's nothing that says "Chapter 1 -- here's the
tree", Chapters 2 through 15, various interesting nodes of the tree,
won't work.  You're doing some kind of taxonomy by defining
"interesting", but 

>  > What I'm unclear on is the necessity of nominating some clade as
>  > "more well-known"; I mean, yes, certainly, a clade that's received
>  > a lot of study based on hundreds of specimens is better known than
>  > one based on two fragmentary specimens and a single paper, but that
>  > is not a stable state of difference.  (Since in principle someone
>  > can always find more specimens and more work can always be done.)
> The Neosauropoda example should clear that up.  Clades based on and
> immediately branching from major branch-points are always interesting.
> A few examples that spring to mind are the node-stem triplets
> {Archosauria, Panaves and Pancrocodylia(*)}, {Dinosauria, Saurischia,
> Ornithischia} and of course {Neosauropoda, Diplodocoidea, Macronaria}.

Well, granted, but there's a lot of people writing papers on marine
annelids, too.  "Interesting" is a wide country.

-- Graydon