Re: New AVES definition refined (more testable?)

["David Marjanovic" <david.marjanovic@gmx.at>]

> HP Mortimer observed:
> <"And at least one of these two characters"?!  That sort of definition
will
> never hold.  We're after shared derived characters, not characters that
have
> a greater probability of being present.>
>
> Please clarify this statement for me.

Not sure if I clarify, but I think the key word is tree-thinking:

> I'm confused because I'm starting from the premise that
> -- whether characters are shared/derived is an inference

yep

> -- that inference is drawn from the characters' presence or absence in
> certain beasts

And on the position of these beasts in the most parsimonious cladogram.

> -- the logic of the inference is based on consistency of the
> presence/absence of characters as linked with other characters.

In short, a parsimony analysis.

> Following these multiple conditional logical arguments (if this... and
> this... and this..., then that... and that...) is tough enough.

That's why it's always a hypothesis :-)
When cladogram A is most parsimonious, then character 1 comes out as a
synapomorphy of clade x, convergently acquired by y, but if B is most
parsimonious...
It's a logical deduction of a testable hypothesis, to be precise.

> Please note that I'm not commenting on the main point of your argument.

Which, if I interpret it correctly, was the question about what we'll do
when we find 2 animals, each of which has one of "these two characters", but
not the same as the other animal -- this would put the monophyly of the
thusly defined group into question. AFAIK this is why the current draft of
the PhyloCode allows only one apomorphy in an apomorphy-based definition.