[Taxacom] PhyloCode & ICZN to "duke it out"?

[Laurent Raty]

Rich,

> Don't think of it as hierarchical, or even as a replacement
> to the Linnaean system.  Think of it as a less ambiguous way to represent
> cladograms in the form of text (names), rather than line-drawings.
> Whereas the Linnaean system was unambiguously NOT designed to represent
> phylogenies, the Phylocode system was explicitly designed for that exact
> purpose.

I can understand the point that the two systems were devised to deal with
two different issues, and may have their own merits when dealing with these
respective issues. What I can't understand however is the suggestion that
the two systems could ever coexist. If we have to live with two systems
acting on the same single set of names, each with a different set of rules,
we will be in the worst possible situation in terms of communicating
anything.

Currently, more and more Linnean names are being "converted" to clades by
giving them a phylogenetic definition - typically a definition that
attempts to make the content of the clade the same as the most widely
accepted content of the Linnean taxon. The basic problem is, this
equivalence of contents will in many cases not survive the phylogenetic
hypotheses under which these definitions were proposed.

Imagine, e.g., that an internal specifier of a "converted" family name is
found to belong to a rather distant family. Everybody agrees on this, and
there are no particular lumper/splitter problems. However :
- As per PhyloCode rules, the "converted" name then makes a big jump
upwards, and ends up covering a larger clade equivalent to, say, a
suborder; we now need to devise a new name if we still want to refer to the
clade made of the former family minus this specifier.
- As per IC*N rules, unless that specifier is the type of the family, it is
simply transferred to the family where it belongs; the same name continues
to apply to the family minus this specifier - a new, more recent name can
in no case become its valid name. (With a bit of bad luck, it's the name of
the family to which the specifier is transferred, that could end up
changed.)

These two systems are fundamentally incompatible. Either one must be chosen
and the other abandoned altogether; or at the very least it would have to
be decided that they regulate two entirely distinct sets of names. Any
other option would unavoidably produce nomenclatural chaos.

Cheers,
Laurent -
--
Laurent Raty
l.raty at skynet.be
Brussels, Belgium