[Taxacom] Excellent post (Paraphyletic species and paraphyletic higher taxa)

[Kenneth Kinman]

Hi Curtis,
      Excellent post.  Thank you for summarising the problems so
clearly.  I agree that a Bessey cactus is sort of like "walking in mud",
because it is not precise enough with all the cladistic knowledge we now
possess for angiosperm relationships.  That is why I don't advocate too
much use of paraphyletic taxa.  If it is not precise, it can often be
authoritarian (which is one reason Hennig's revolution became so
popular).    
      HOWEVER, this is no excuse for rejecting precisely-defined, formal
paraphyletic taxa.  Hennig went too far and rejected such taxa
altogether, thus creating a new form of authoritarianism (the worst
manifestation of which will be the PhyloCode).         
      Paraphyetic taxa can be (and should be) precisely
defined/delineated (which can then be scientifically challenged and
modified if necessary).  As you noted, with clades you can say "show me
your synapomorphies".  But there is no good reason that an occasional
important paraphyletic taxon cannot be precisely defined (a definition
which can then be challenged and modified if needed).  You do this by
giving the synapomorphies (as with clades), but additionally giving the
synapomorphies of the autophyletic exgroup (but which the paraphyletic
taxon lacks).  For example one could give synapomorphies for a
paraphyletic species (brown bears), but you  also need to list the
synapomorphies of the exgroup (polar bears), which brown bears lack.            
      This would produce classifications which would be a common ground
for debates between those who advocate very few such paraphyletic
groupings and those who advocate more.  There would still be healthy
debates between the two camps, but it would end the war (and minimal
cooperation) we now have due to authoritarianism on both sides.  Too
many (or poorly defined) paraphyletic taxa produced too much
"traditional" authoritarism, but that is no reason to swing the pendulum
to Hennigian authoritarianism (no formal paraphyletic taxa at all).  A
healthier middle ground is badly needed and very LONG overdue. It would
not only be healthier and more cooperative, but also make
classifications more stable and useful.             
        ----------Ken Kinman                
-------------------------------------------------------------

Curtis Clark wrote:
For me, cladistics is a tool, not a religion. Grouping taxa by shared
apomorphy has an elegance that I never found in my pre-cladistic
systematics (a student in one of my classes once likened the Bessey
cactus to "walking in mud"). In that sense, it's like a torque wrench:
elegant at torquing bolts, but not suitable for drying oneself off after
a shower. 
Verne Grant predicted in the second edition of Plant Speciation that
hybridization in the flowering plants was so pervasive that phylogeny in
that group would never be determined. And he didn't mean Hennig
phylogeny, he meant Simpson phylogeny. IMO he has been proven wrong
(depending, of course, on your meaning of "determine"): large parts of
angiosperm phylogeny are well-supported by multiple lines of
morphological, chromosome structural, and sequence data. 
As a tool, grouping by shared apomorphy pretty much always works (if
you're right about the apomorphies). Assuming that groups so diagnosed
are non-overlapping doesn't have as good a track record: hybrid
speciation and lateral gene transfer (by introgression or more
"esoteric" mechanisms) does muddy the waters. Cladistics provides *a*
tool (certainly not the only tool) to explore that. It provides a null
hypothesis ("the distribution of character states in this group is the
result of cladogenesis alone") that can be falsified. 
There are plenty of documented cases of hybrid speciation involving
parent species that are somewhat distantly related. We can assume that
at least some extant lineages arose from hybrid species such as those.
We recognize the possibility because of character incongruence. But
character incongruence has many sources, a major one being homoplasy:
calling something a homology that isn't. Distinguishing incongruent
homology (from hybridization) from incongruent homoplasy (from bad
interpretation) is non-trivial. 
Let's say that we had established with a high degree of confidence that
a named lineage originated from a hybrid between distantly related
species in another named lineage. Would I accept the progenitor lineage
as a paraphyletic taxon? Perhaps. I don't object in principle. Did Aves
arise from Reptilia by hybrid speciation? I've never read of that
hypothesis. And there's the rub. There is no single criterion for
erecting paraphyletic groups. They can be based on evolutionary grades
(I posited in a previous Taxacom post a rationale for an entirely
grade-based taxonomy, but no one seemed interested). They can be based
on presumed tokogeny. They can be based on morphological gaps that are
subsequently narrowed by discovery of new species. They can be based on
whim. And not even Ken's notation explains which of these or other
reasons are used. 
With a clade, you can always say "show me your synapomorphies". A
paraphyletic group is argument by authority. 
And my biggest reason for rejecting paraphyletic groups might seem
trivial: they don't play well with others. Any classification that
accepts a paraphyletic Reptilia from which sprang a monophyletic Aves
will never contain a monophyletic Saurischia or Archosauria. The only
way to have those taxa would then be to develop an entirely new scheme
of nomenclature--we might call it "Phylocode". I'd rather see our
existing system use clades.