RE: Philosophies for Character Ordering

["Jaime A. Headden" <qilongia@yahoo.com>]

Mike Milbocker (mmilbocker@psdllc.com) wrote:

<I'm not sure of the definition of "correctly" here. I do not believe it
is consistent with the theoretical basis of cladistic systematics to have
states in a character that are not independent from one another and
mutually exclusive. If an organism can be in both states 0 and 2, then the
character is ill-defined.>

  Mickey was correct in the sense that in a logical system, progression
from 1 to 2, then to 3, must allow the animal to go through condition 2;
if the animal has condition 3, the assumption of an ordered character
asserts that it must go through 2 in that theory. This is just as
applicable and is not assumed to be the "right," but is a "logical,"
argument.

<The machine appears to be making alot of assumptions here. Why not make
the characters dependent on one another, if one is willing to allow the
states to be dependent. For example, write a character that says if
Character A is in state 0 and Character B is in state 1 then Character C
is in state 0, otherwise state 1. If the states of Characters A and B map
uniquely to the states in Character C, you can discard A and B without
loss of information.>

  As Mike noted before, the person selecting which characters as ordered
versus unordered is making the assumption, not the machine. The machine
interprets this series of transformations as being an algorithm, not that
it assumes anything.

  As for the co-dependant character system, that too is an assumption of
the relation of two characters. Treating some characters as a suite, or
complex, may over-bias a single character. This can be done now my making
each complex it's own character, or making them each a state of a "suite
character." Virtually everything that can be made as a descision in the
programming language is an assumption on the part of the analyzer, from
formulation of the character, coding of the character, which taxa to use,
number of characters, relation of ordered versus unordered, how many
should be one or the other, using step-matrices, enforcing trees (e.g.,
force *Heterodontosaurus* to be a part of Ornithopoda sensu stricto,
rather than sister to Marginocephalia).

<Can you please state the intrinsic assumption and associated corollary,
if an assumption can have a corollary.>

  Say that two characters afford that two taxa are more similar to one
another than any other relationship. A third character they share is known
among more taxa. This third character is assumed to be a synapomorphy of
the clade even if known in other taxa, because it diagnoses an ingroup,
especially if the produced relationship has a sister taxon that lacks this
condition. By this corollary, this third character should not be treated
as if it were assumed to belong to any particular suite of transformation,
in which case its broad usage becomes limited. It would then argue that
that taxon also had the previous condition, and gives a character to a
taxon for which it does not possess. In fact, the utility of multistates
in this case is incredibly reduced unless the states are unordered,
because variable shapes of thyreophoran cranial armor, the shape of the
deltopectoral crest, etc, become increasingly assumptive of their unknown
primitive relationship. This relationship can be determined as if all
characters were unordered and the tree showed that the character
acquisition was in a progressive line, armor was less variable, etc., and
without ordering. Ordering LIMITS tree generation, which I think it one
reason some use it, when using MacClade or PAUP*.

  Cheers,

=====
Jaime A. Headden

  Little steps are often the hardest to take.  We are too used to making leaps 
in the face of adversity, that a simple skip is so hard to do.  We should all 
learn to walk soft, walk small, see the world around us rather than zoom by it.

"Innocent, unbiased observation is a myth." --- P.B. Medawar (1969)


        
                
__________________________________
Do you Yahoo!?
Yahoo! Domains ? Claim yours for only $14.70/year
http://smallbusiness.promotions.yahoo.com/offer