parsimony/biology (was: Striking a balance...)

[Thomas DiBenedetto]

-----Original Message-----
From: Pierre Deleporte [mailto:Pierre.Deleporte at UNIV-RENNES1.FR]
>...how can you pretend that the choice of parsimony as
>applied in MP has no biological implication?

I can assure you Pierre, I am not pretending.
As I tried to explain yesterday in my respnse to Kirk, all of the biological
and evolutionary assumptions that are part of a phylogenetic reconstruction
effort enter at an earlier stage than the point at which we choose and
implement the parsimony criterion.
They enter at the point when we define a character and hypothesize that it
is a homology. We note for example, that humans and kangaroos and the
platypus all have hair. We define the character, and code it the same for
these taxa, as opposed to the condition we code in reptiles. We hypothesize
that hair is a homology - homology being generally understood as the same
character in different taxa present as a result of descent from a common
ancestor. This homology hypothesis is inherently a hypothesis that the human
species, the kangaroo species and the platypus species all derive from a
common ancestral species. It is at once a phylogenetic hypothesis as well as
an origin-of character hypothesis. That is what homology means.
At this point, all of the necessary biological and evolutionary assumptions
are in place. We have a grouping hypothesis, a simple phylogenetic
hypothesis - Mammalia is a node. As we define and code additonal characters,
we are hypothesizing additional nodes. When our matrix is complete, we have
a set of x number of hypothesized nodes, x number of simple phylogenetic
hypotheses. At this point, we select and implement the parsimony criterion,
and it does nothing except sort through the hypotheses, in a logical,
mathematical procedure to select the overall topology which is made up of
the nodes that are most supported by the characters. There are no additional
biological or evolutionary assumptions added.
It is not necessary to use the parsimony criterion of course, one can do
whatever one chooses. But the justification for using parsimony is nothing
but the same justification that one uses in every other (non-phylogenetic)
instance in which we use parsimony - a logical criterion that constrains us
to select the result that most efficiently summarizes the evidence - not to
postulate complexity unnecessarily. The phylogenetic use of parsimony
requires us only to select the tree that has the most support in the
evidence. That is logic, not biology.
-tom