Humans and apes

[A.P. Jason de Koning]

John:
> As a phenetic enterprise it happens to be consistent with itself
> ...
> The data may be analyzed using cladistic algorithms, but the data (the
> characters) themselves are not identified as cladistic in the first place
> (i.e. the sequences are not sorted into hypothesized synapomorphies that
> can be tested against each other - at least that is my current
> understanding. If I am wrong I welcome correction.

So how are you proposing that phylogenetic trees are constructed from
morphological data??

I think you are wrong.  The point of doing maximum likelihood is to fully
generalize the cladistic enterprise, but to maximize the use of all
available information.  I don't see how sorting characters before you
analyse them has any effect on the outcome.  In ML, you in effect propose
all possible synapomorphies (for all characters), and evaluate them based on
their joint probability.  More correctly, you propose all possible series of
changes that lead to the observed data (which include the synapomorphies),
and select the tree that is most likely reponsible for the most probable
evolutionary paths (all evolutionary paths are considered).

I think we are not talking the same language here.  That "ain't my momma's
definition" of a phenetic enterprise...

> >But since phenotypic traits are
> >moulded more directly and (probably) frequently by natural selection,
they
> >become more rapidly skewed away from the underlying historical signal.
>
> This is just a theoretical proposition. Whether it is true is another
matter.

Are you arguing that morphological traits commonly evolve independently of
fitness? (see my last reply to Brian)

> Thanks. Wherever this leads its interesting and I will learn from the
exercise.

I'll be interested to see your final exhibit.  Maybe I'll come by if I'm in
the area :)

Cheers,
- Jason