Positivism vs Realism
James Francis Lyons-Weiler
weiler at ERS.UNR.EDU
Fri Dec 12 09:07:22 CST 1997
On Fri, 12 Dec 1997, Thomas Pape wrote:
> >[JL-W:] When there is evidence that the explanation needed is
> >more complex than parsimony would allow.
> This seems to me to be 'twisting' the concept of parsimony. Parsimony as a
> tool does not discard "evidence", only ad-hoc explanations.
I'm not sure what you mean here - my answer to question
didn't involve discarding evidence - only missing it.
> Which is in agreement with our observations.
In a general way, but the expectation need not
obtain in any paritcular case.
> > My comment should have been more obviously directed at
> > inheritance at the lineage level - where the diconnect
> > between geneaology and cladogenesis exists. So, yes,
> > lineages can and are expected to aquire new states that
> > are independent of their ancestors,...
> This may be beside the point and not quite true. ANY character state should
> be supposed to have evolved from some other (per definition more ancestral)
> character state. Character states are the properties of lineages and
> lineages have ancestors. Even in the case of horizontal transfer and
> reticulate genealogies (hybridising), we do not interrupt the 'descent with
> modification' nor create ancestry-independent states. Congruency should
> therefore be our _a priori_ assumption. That it may be difficult to
> recognize true lineages is another matter.
Once state B has arisen and state A replaces it in a lineage,
A is autapomorphic to the lineage and the information is
reflects is, for some data, independent of the information
carried by B. When I say "independent" I don't mean that
the chain of descent is broken, only that the information
carried is independent. A subtle point missed by some is
that the models of descent with modification does allow
for all sorts of non-hierarchical processes - but parsimony
Whenever congruence is expected, and used a the criterion,
any level of congruence is acceptable. A better posed
question would be "how much congruence have we found
beytond that which we would expect by chance alone?"
There is still a lot of room for the development of
such tests - and a definite need for good tests like this.
The field is struggling with what constitutes a good test
based on inferential statistics - and part of the slow
progress is the potential misleading effects of optimal
trees when they or their characteristics are part of the
test. Hence tree-independent tests...
> > Parsimony clearly assumes [a] model of
> > lots of descent with little modification...
> I have no problem with that -- this is exactly what we observe: "lots of
> descent with little modification". Difficult to imagine reconstructing
> evolution if we had it the other way round. :)
My point exactly - we can either expect congruence and
convince ourselves that phylogenetic estimation
is an easy task, or we can be more skeptical and demand
critical tests of the level of congruence found in
the data, and, in the process, in many a case, learn much
more about the organisms and their attributes that we
would have missed with parsimony alone.
More information about the Taxacom