[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
Re: Darwinopterus notes
2009/10/17 David Marjanovic <email@example.com>:
> And a lot more characters! 117 characters for 56 taxa -- that's less than
> twice as many. Under three times I wouldn't bother publishing.
Why? Because it tends to show less resolution or because you suspect
they have not searched enough characters given the appearent diversity
they encompass in the analysis?
> Also, all characters were unordered. Some of them are continuous, like
> number 6: "Rostral index: 1.5 or less (0); 1.5-3.0 (1); >3.0 (2)". This is
> not defensible (Wiens, 2001, Syst. Biol.) -- the assumption that underlies
> ordering, namely that it is easier to change from a character state to a
> similar one than to a more different one, was _already_ used to divide the
> continuous character into states; if it's easier to go from 1.3 to 1.4 than
> to 1.6, it must also be easier to go from 1.3 to 2.0 than to 3.5, in other
> words, not ordering the character is logically inconsistent.
Thanks you this time for the citation! I was always convinced that
ordering was better, and never read a defense of non-ordering (I know,
shame on me). I though perhaps the unordering guys think considering
it is easier for 1 to go to 1.5 than to 2.3 implies relying on a
supposition on how evolution works (i.e., that gradual changes are
more likely to occur than abrupt ones).
Although Farris supported the use of ordered characters (indeed, these
characters were first discussed by Farris as far as I know, and their
original name is "additive" instead of "ordered"), he also called for
not reconstructing phylogenies on the basis of preconceived ideas on
how evolution works (this way attacking likelihood, the models on
which this approach is based, and Dollo parsimony).
May have the unordering guys used this to justify their choice?