Striking a balance, weighting and Cladistics

[Thomas DiBenedetto]

-----Original Message-----
From: Kirk Fitzhugh [mailto:kfitzhug at NHM.ORG]
>...each homology hypothesis cannot stand as "evidence" for a
>more inclusive explanatory hypothesis in the form of a cladogram. The
>"evidence" for inferring cladograms are shared similarities which are in
>need of explanation, in conjunction with some causal theory. Hence, the
>use of the parsimony criterion is not simply a "logical tool," but
>represents the application of theory to observations (i.e., effects) that
>are in need of explanation. To deny the place of theory in this
>inferential process is to deny the ability to explain.

Kirk, Let me try to explain why it is that I disagree with this perspective.
In the process of doing so, perhaps I can clarify for Mark Garland the
relationship that I see between character homology and species
relationships.
As I see it, each character in a matrix is not only a hypothesis about the
historical relationship of the character itself, but of the organisms and
species which bear the character. To say: "these proteinaceous outgrowths in
these different specimens are all the same, to be given a common name - hair
- because we hypothesize that they all are related by being descendant from
a common ancestral character" inherintly states as well that "these hairy
specimens are all descended from a common ancestor". Each character in a
matrix is thus a simple phylogenetic hypothesis as well as a character
homology hypothesis. The cladogram is simply a compendium of these simple
phylogenetic hypotheses, combined in such a way (i.e. parsimoniously) such
that the preferred result is the topology that minimizes the number of
simple phylogenetic hypotheses that need be rejected in whole or in part.
The explanation, or the reference to causal theory, enters at the point that
one makes the hypothesis in the first place. The notions that life evolves,
that characters descend, or whatever other theory is necessary, are
referenced when we seek to place the characters in a historical context.
Saying that two species share a common ancestor is grounded in the exact
same theory that allows one to say that the character state found in
organisms of two species are homologies. In other words, nothing need be
added, in a theoretical sense, to go from character homology hypothesis to
species-phylogeny hypothesis. Put another way, if you can explain homology,
you can explain phylogeny.
And so the use of the parsimony criterion need not be grounded in any causal
theory - it serves merely to logically combine simple phylogenetic
hypotheses into a comprehensive hypothesis in a manner that least
contradicts the simple hypotheses.
-tom