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

Re: something's wrong here: Qianosuchus phylogeny



On 9/4/07, Mike Taylor <mike@indexdata.com> wrote:
>
>  > (See, this is why higher-level paraphyletic taxa are a bad idea....)
>
> Well, no ... you've not shown that higher-level paraphyletic taxa are
> a bad idea, you've just shown one possible consequence (that one taxon
> can be located inside another) and _asserted_ that it's bad.

One taxon can be located inside another in phylogenetic taxonomy, too,
so that's not the problem. This is the problem: ancestor A gives rise
to independent lineages B and C. If you only allow monophyly, then
there is only one possible way to divide them into higher-level
groups: a parent clade, ABC, and two subclades, B, and C. The taxonomy
proceeds directly from the phylogeny--it is basically identical to it.
But if you allow paraphyly, you have to choose between several
systems. You could recognize a group, AB, which gave rise to C. Or you
could recognize a group, AC, which gave rise to B. Which option is
better? Objectively, neither.

Obviously, you could have a system where AB and AC are both valid
taxa. This is not possible under the rank-based codes, but, in theory,
you could do it. But this opens up a huge number of taxa that can be
named, an order of magnitude greater than the number of clades that
can be named (which is already overwhelming). Would it really be
useful to have names for all of these, when they can easily be
referred to in terms of clades, or even using an occasional
well-selected informal name ("stem-mammal", "Triassic dinosauromorph",
"fish", etc.).

And the lack of a consistent hierarchy would breed confusion. I could
legally recognize Phytodinosauria and Saurischia
simultaneously--sauropodomorphs (and probably herrerasaurids, etc.)
would be members of both. And I could use both whether I followed
Bakker's phylogeny or the more widely-accepted ones. Not that
overlapping taxa are necessarily a bad thing, but shouldn't the
overlap be related to some real phylogenetic phenomenon (e.g.,
hybridization)? Shouldn't the nomenclature actually communicate
something about the phylogeny, rather than wishy-washily adapting to
fit just about any tree?

So, in short, allowing formal names for paraphyletic groups (at least,
above the population level) introduces either subjectivity, if you
disallow overlapping groups, or a huge, generally unproductive mess,
if you allow overlapping groups. Maybe there's a more workable
solution for allowing overlapping groups, but I've yet to see it.
-- 
Mike Keesey