Striking a balance, weighting and Cladistics

[Kirk Fitzhugh]

Tom,
Thanks much for the clarification in your message of 2/21. I've not yet
read the comments of others posted today on this topic, so please excuse
any redundancy. Since we both seem to agree that cladograms satisfy our
goal as scientists to causally account for what we observe in the world,
I'd like to comment on some of your statements which I think are logically
inconsistent with that goal.

You suggested that a character in a matrix contains both an hypothesis of
phylogeny as well as an hypothesis of homology. These are the same thing.
Each homology hypothesis is a phylogenetically-based explanation: "the
observed individuals share the property of hairiness due to common
ancestry." In your example, you associate the naming of properties as the
result of a need to explain the existence of the shared properties: "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." From an
epistemological standpoint, the naming of properties simply denotes our
perceptual beliefs that what we observe are identical in some sense. It is
only after the application of such names that we can ask the question as to
why those properties exist among some set of individuals. This makes the
distinction as originally intended by Owen between homologues and homology.
Owen was very clear that structures given the same name are to be called
homologues, whereas the explanation of those similarities is by way of
homology (regardless of by way of archetypes or common ancestors).

This would then mean that one *might* be able to say that a data matrix is
the totality of homology hypotheses, i.e., individual explanatory
hypotheses, but this does not work for two reasons. First, if we regard a
cladogram as a causal explanation itself for the presence of shared
similarities - which must be the case if one is to be allowed to say that
internal nodes represent speciation events, and internal branches
hypothesized common ancestors - then this explanation must be in answer to
a set of specifiable questions. There must be distinct relations between
effects that are observed, the why-questions relating to those effects, and
an explanation(s) that offer answers to those questions. As a result, a
data matrix actually is not a set of homology hypotheses, but represents
the set of why-questions pertaining to the properties we observe. If, on
the other hand, a cladogram is nothing but summation, then there is no need
for cladograms since the goal of science is to explain, and this can be
done with homology hypotheses. This then is where the second reason comes
in as to why a data matrix is not a set of homology hypotheses. The basis
for treating all observations in the same explanatory context is because of
the requirement of total evidence. Unfortunately, the requirement has not
yet been properly articulated in the cladistic literature, pro or con. If
we wish to claim that cladograms are nothing but the grand explanatory
summation of individual explanatory (homology) hypotheses, then the
requirement of total evidence is irrelevant. That the requirement has been
regarded as fundamental is that it pertains to the causal relevance of some
effects on others. If to causally account for some set of effects has
relevance to the inference of other causal accounts, then all relevant
effects must be considered in the inference of cause. What this then
entails is that while homology hypotheses each are themselves separate
explanations, the requirement of total evidence stipulates that these
hypotheses be set aside and our why-questions all answered simultaneously,
which would be in the form of a cladogram.

I completely agree with you that causal theory does enter the picture at
the point one posits homology, but which is prior to the designation of
homologues, i.e., naming. The point where I think we differ is that I
contend that cladograms are explanations that use the same causal theory as
homology, but have replaced individual homology hypotheses in order to
satisfy the requirement of total evidence. In both instances of inferring
homology and cladograms, parsimony is invoked because of the conjunction of
effects and causal theory in the inference of both. The only difference is
that the premises differ in scope in each, but each uses the same causal
theory of common ancestry (but if not, which is the case with maximum
likelihood, then there are real philosophical problems). With cladograms
regarded as formal explanations, then instances of homoplasy are to be seen
as explanatory parts that account for shared similarities by way other than
common ancestry, which I think is consistent with what Farris (1983) said.

Best,
Kirk


"True science is distinctively the study of useless things. For the
useful things will get studied without the aid of scientific men.
To employ these rare minds on such work is like running a steam
engine by burning diamonds."

Charles Sanders Peirce

------------------------------------------
Kirk Fitzhugh, Ph.D.
Associate Curator of Polychaetes
Research & Collections Branch
Los Angeles County Museum of Natural History
900 Exposition Blvd
Los Angeles CA 90007
Phone:   213-763-3233
FAX:     213-746-2999
e-mail:  fitzhugh at bcf.usc.edu
------------------------------------------