Re: Strange thoughts on PN - was Re: BAD vs. BADD

[David Marjanovic <david.marjanovic@gmx.at>]

----- Original Message -----
From: "Martin Baeker" <martin.baeker@tu-bs.de>
Sent: Monday, August 21, 2006 10:00 AM

> Sure, but we don't usually talk about sparrows, falcons, ducks etc.,
> at least not for about 90% of our time, as we should according to the
> number of species.

Bah! Those are all the same. No morphological, ecological, or temporal 
disparity whatsoever! (In first approximation anyway.) Here we concentrate 
on the interesting ones. =8-)

> Everybody here agrees that group A=(Megalos.+Iguanodon) is useful, as
> is B=(Passer+Archaeopteryx) - where () denotes MRCA - despite the fact
> that these are also quite arbitrary - there will be a critter looking
> *almost* like the first member of A that is not a member of it (like
> Lagosuchus), but still, the concepts are useful, albeit arbitrary.

- Apomorphy-based definitions are allowed.
- Arbitrary, yes, but only in one direction. In rank-based classifications, 
when you decide where the birds begin, you are at the same time inevitably 
deciding where the dinosaurs stop. Phylogenetic nomenclature allows you to 
have one without the other -- having your cake and eating it, too.

> Why is then A\B (A without-B) such a big no-no? From the frequency we
> hear "non-avian dinosaur", this gouping is *useful* for many purposes,
> except, of course, when discussing phylogeny, so, what is *useful* may
> get a name, at least a vernacular name.

You can give it one. You can even give it a phylogenetic definition: "clade 
A without clade B" for example. All you won't be allowed to do under the 
PhyloCode is having its name registered.

Think of it: in a Linnaean classification you can't have a taxon for the 
egg-laying theropsids/synapids, unless you abolish Mammalia. Phylogenetic 
nomenclature is _better_ suited for the discussion of paraphyletic groups 
than Linnaean nomenclature is!

> > > Thought experiment: [...]
> >
> > That's about species concepts, not about phylogenetic nomenclature ( = 
> > how to
> > name clades).
>
> Doesn't really matter - the thought experiment was meant more as an
> analogy. So, if archaeopteryx had been wiped out and no birds had ever
> evolved, what we now call non-avian dinosaurs (B/A, see above) would
> be considered a useful and allowed group, but because archie did
> survive, it is not?

Dinosaurs excluding just *Archaeopteryx* are not monophyletic either...

> Don't get me wrong, I'm all for cladistics when it comes to revealing
> phylogenetics.

You are still getting me wrong: phylogenetic nomenclature is entirely 
independent from cladistics. :-)

> Who knows, 100 years from now, we may be able to quantify
> morphological differences in some way, and then we can define
> Hominidae as everything that does not differ by more than 20
> Morphological Units from homo sapiens, and dinosaurs as anything that
> does not differ more than 100 MU from Iguanodon, and, thus, somewhere
> in theropoda, there would be a boundary line and everything above it
> would not be a dinosaur anymore. Sure, that's also arbitrary
> (boundaries would shift if we switched 100 to 105), but all groupings
> are.

In this paragraph you have reinvented phenetics, aka Numerical Taxonomy. So 
let me then recapitulate its rise and fall:

In the 1950s some people got fed up with all the splitting and lumping and 
the swapping of taxa every time a new phylogenetic hypothesis -- or rather 
speculation -- was proposed. There had to be some objective criterion that 
could be used for classification, they thought. So they invented phenetics: 
All characters should be measured, without concern for how "important" or 
"distinctive" or "interesting" they might be, and entered into a data 
matrix; from this matrix a computer calculates a distance matrix (consisting 
of the distances between any two pairs of OTUs -- operational taxonomic 
units -- they invented that term), and from this a tree-shaped diagram can 
be calculated for making the whole thing easier to grasp, without (OTUs that 
are more similar to each other are separated by fewer nodes) or with branch 
lengths (OTUs that are more similar to each other are separated by more 
total branch length). Some proponents, I hear (not having read the papers, 
this is all mythology to me), were radical: if the whales cluster among the 
fish, then let them be fish, at least they will stably be so. The Linnaean 
ranks could at long last be defined (as percentages of similarity).

Some 20 years later phenetics was just about extinct (surviving only among 
molecular phylogeneticists, who merrily kept making phenograms for decades 
and honestly believed they were making cladograms; this practice still 
hasn't quite stopped). It seems that there were two reasons:

One was that there simply was no end to it. "Character" cannot be defined 
(unlike "phylogenetically informative character"!!!). Thus, the decision 
what to put into the data matrix and what not is subjective, turning 
phenetics into a complete failure measured by its main goal.

The other was that people actually _want_ to talk about phylogeny, about 
evolution. And this is why phenetics was replaced by cladistics and not by 
"evolutionary systematics".

There may be yet another good reason ("it's so newfangled" not being a good 
one): The best way to represent the results of phenetic research is not a 
tree or any other hierarchy (such as a Linnaean classification). It is the 
distance matrix itself. In other words, phenetics would in reality be 
completely impractical, even if there were an objective definition of 
"character".

Ironically, phenetics has kick-started cladistic practice. The use of 
computers, the term "data matrix", and the term "OTU" all sound familiar, 
and rightly so. Hennig had calculated his MPTs by hand; the limits on the 
size of the data matrix, and the probability of missing at least some of the 
MPTs, are obvious.