AW: Notarium question

[Mike Taylor <mike@indexdata.com>]

evelyn sobielski writes:
 > > If you score them differently on the assumption that they are
 > > separate apomorphies, then you're assuming what you're setting
 > > out to prove.
 > 
 > On the other hand, if you score simply "notarium present (y/n)?",
 > you may not be able to show they are separate apomorphies *even if
 > they were* because your scoring assumes that the notarium is a
 > monolithic entity - which any cladistic analysis will try to
 > optimize towards a single origin.

That is exactly what you want: the analysis will balance the
desirability of parsimony of the notatium (single origin, no
reversals) with that of all the other characters.  Like I said, you do
the analysis to _discover_ whether this (and other characters) have a
single origin or multiple origins.

 > (This is the mistake that led Cracraft to "prove" that
 > Hesperornithes, loons and grebes are a clade to the exclusion of
 > all other birds living and extinct. In the case of the notarium,
 > its presence was used to argue for the monophyly of "Metaves" -
 > which, as it seems now, are almost certainly a pseudoclade. For a
 > similar error, consider the fallacy "everything with feathery
 > integument is a member of Aves".)

I am not familiar with Cracraft's work, but it sound like his problem
was not an a priori assumption that the notarium has a single origin,
but over-emphasizing one "important" character to the neglect of
others that tell a different story.

And notice anyway that assuming a single origin of the notatium, as
you say Cracraft did, is a _completely_ different thing from scoring
the presence of a notarium as a single character and leaving it up to
the analysis to determine how many origins that character had in the
most parsimonious trees.  That's the whole point: we try not to assume
_anything_: just make observations, and let the computer turn its
wheels and tell us what those observations indicate about the
evolutionary sequence.

 > This need not yield "better" results (i.e. better-supported and
 > less polytomies). But uncertainty that is correct is preferrable to
 > certainty that is wrong.

No argument there!

 _/|_    ___________________________________________________________________
/o ) \/  Mike Taylor    <mike@indexdata.com>    http://www.miketaylor.org.uk
)_v__/\  "These lions you saw.  Did they ... eat ants?" / "Yes, that's
         right." / "No, those weren't lions you saw.  They were anteaters"
         -- Monty Python's Flying Circus.