paraphyly (still)

Tom DiBenedetto tdib at UMICH.EDU
Fri Jun 13 09:29:49 CDT 1997

On Fri, 13 Jun 1997 10:13:48 EDT, Robin Panza wrote:

>This statement is based on one of the primary underlying tenets in
>cladistics, that evolutionary rate is more or less constant at the molecular
>level (to be able to assume that degree of divergence measures age of

I have never heard a cladist make that claim and I cant imagine how
this can be seen as consistent with cladistics, at least as I
understand it. Cladists explicitly do NOT assume that the degree of
divergence need be correlated with age of divergence, This is the
essential point in the very heated antagonism between cladists and
pheneticists; that the degree of divergence, whether in overall
phenotypic similarity of some measure of molecular distance, is not
to be expected to accurately reveal relationships. Cladists rely
rather on distinguishing between three causes for percieved
similarity; apomorphic similarity (the character is a homology
evolved in the common ancestor of the species considered),
plesiomorphic similarity (the character is a homology evolved in a
common ancestor of a more inclusive group than that under
consideration), or convergent similarity (the character is not a
        The operative assumption is that homologies are expected
(almost by definition) to exhibit a congruent pattern, especially
when all aspects of the phenotype and genotype are considered. Thus
the search for congruent pattern in as wide a character-sampling
regime as possible, can be expected to reveal the traces of the only
causal factor which we think can impose patterned similarity on
potentially all characters; history.
        All cladists that I know of are very well aware of the fact
that evolutionary rates are not constant, and it is cladists who
ususally emphasize that uncomfortable fact to those who truly do rely
on some regularity of evolutionary rate (e.g. many in the school of
statistical phylogenetics). It is by distinguishing between the
presumed causes of similarity in individual characters, by seeing the
extent to which they fit into an overall pattern of character
distribution, that we hope to uncover the traces of history.

> If a taxon has a subgroup that
>changes (far enough to call another species or *any* other rank) while another
>subgroup is still recognizably the same as it was, then we have "budding off",
>not the "birth of a new hierarchical rank".

It seems to me that a "rank" refers to the place in a hierarchy. If
the hierarchy evolves, in the sense of becoming more complex, then
new levels ("ranks", if you wish to acknowledge and recognize
reality) are formed.

 Traditional (old-fashioned,
>non-scientific, or whatever name you choose to use) taxonomy has no problem
>accommodating this possibility; cladistic theory has not (yet?) become robust
>enough to deal with it "scientifically".

Traditional taxonomy "deals" with this by assigning equal rank to
entities that are not symmetrical within the historical hierarchy.
This can be defended by appeals to tradition, stability, and
practicality perhaps, by I cant see how you can argue that it is a
more accurate reflection of the actual braching pattern.

Tom DiBenedetto       
Fish Division                                   tdib at
University of Michigan Museum of Zoology

More information about the Taxacom mailing list