On Mon, 5 Feb 1996, Michael Hobbs wrote:

> I think the real problem is that the branches are not so real. Where
> the branches are obvious, we haven't had problems of lumping and
> grouping and changing of classifications.

The branches *are* real, but it can be difficult at some of the branch
tips to tell whether there is one big, fuzzy tip, or two or more distinct
tips ("species") that are difficult to distinguish. As you move away from
the tips, this ambiguity disappears. This is why -- ironically enough --
"higher" taxa (clades, not categories) are less subjective than species.

[small snip]

> We tend to name the extremities. Sometimes, after scientific inquiry,
> these are determined to be only morphs (i.e. dark morph and light
> morph). Sometimes these are determined to be subspecies/races.
> Sometimes these are species within a genus. Sometimes, what appears at
> first to be a single population turns out to be two which are not even
> related, but which have converged in morphology; look-alike
> butterflies come to mind. Sometimes there is a continuum which
> confounds our desire to pigeonhole.
^^^^^^^^^^^^^^^^^^^^
This is one of the problems with categories. What if we simply named the
branches and didn't assign them to pigeonholes (categories -- order,
family, genus, etc.)? If what we originally thought to be a single
lineage (A) is later recognized as two, we don't need to "split" A into
two taxa of that each have the *rank* that A once had; rather A still
exists as a taxon, but now it has two subtaxa (B and C) included within it.

>[...]

Re the examples from mallows and Townsend's and Hermit warblers (snipped
here for brevity), whether the populations/groups of individuals in
question are species or parts of species all depends on what you think
a "species" should be.

> So I think it's safe to say that there will always be lumping a
> splitting, regardless of how we choose to name what we see.

Not in a system that doesn't pigeonhole taxa into categories (family,
genus, etc.). There may be subdivision, but that's not the same as
splitting, because the newly recognized taxa don't assume the
*categorical rank* of the original (there is none); -- they are simply
nested within the original.

> Publication of the current standard allows scientific communication
> without confusion.

But when there are divergent hypotheses about the status of a population
as a "species" or "subspecies," or whether vultures are more closely
related to storks or to hawks, then there isn't a "current standard."
*Choosing* hypotheses via committee and making them de facto standards,
thus excusing the majority of ornithologists from the chore of "keeping
up with taxonomy," and giving them the blessing of the AOU to ignore
alternatives, does not serve science, it serves convenience (and as Dennis
noted, this committee approach is apparently unique to ornithology).
Scientific societies should not be in the business of endorsing
hypotheses with the equivalent of a "Good Housekeeping Seal of Approval."
Re Dennis's question as to a better way of doing it, I'd say the less
convenient, but more honest way everyone else does it -- by leaving
disagreement over systematic hypotheses in full view.

David Wright
dwright at u.washington.edu