[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
Regarding characters, their proliferation, and their objectivity:
I am personally very comfortable with the large numbers of
characters being put forth in the literature. Although the sheer volume of
data is daunting for non-specialists (and sometimes for specialists,
especially victims of the Holtzicon, I imagine!), I find that, being
familiar with a group allows one to quickly and easily evaluate
characters. However, as Josh points out, this is highly dependant upon the
degree to which the author has attempted to convey their observation in a
cogent manner. In hadrosaurs, many authors have phrased characters in very
obscrue ways, by referring to structures by obscure terminology or, as
often, non-obscure terminology described in obscure journals or in a
footnote of a subparagraph of a 200 page document.
So I do agree that characters need to be phrased such that they
may be evaulated objectively. However, ther seems to be a need on the part
of many people, especially non-cladists, to see some "meaning" in a
character. People get uppity if your character seems to be a random
observation which does not obviously relate to morphology ("who cares
whether the humeral condyles are more than 25% the length of the bone?). One
thing to keep in mind is that a character is an OBSERVATION of potential
homology, not a rigorous argument. We use parsimony to evaluate the
character, and then interpret its meaning or significance.
I use ratios, etc. as a means of making an observation
reproduceable (in the hypothetical humeral condyle example, this is one
way to describe a "robust" humerus). This does not necessarily mean that I
attach special meaning to this ratio, simply that I have observed a
potential homology, and I want other people to be able to evaluate it. I
would phrase this something like, "Humerus: 0) gracile, 1) robust
(lateral width of humeral condyles greater than 25% of the length of the
humeral shaft). There are three ways I can think of offhand to attack this
character: 1) show that there is nothing at all special about the 25%
cut-off, and most of the ingroup+outgroup humeri form a close, continuous
spectrum across it, and suggest a different metric for scoring the
character, 2) argue, from a multivariate perspective, that there really is
no discrete difference in ingroup humeri, 3) find a discrete character
whichis 100% correlated with the 25% cut-off, and use it instead (and
smack me for not noticing it before).
To my knowledge, there is no means of mechanically reducing
specimens to objectively meaningful characters. Character identification
is a manual process, and is therefore somewhat subjective. This does not
damn the method, because characters can be evaulated and argued on
objective grounds (i.e., is the morphology truly present, is this a
manifestation of another character which is already coded, is this
character polarized correctly). I believe many would agree that this
subjectivity does not give one the right to reject characters one does not
feel are "meaningful" a priori of analysis. However, character
argumentation should be a process of describing and codifying
observations, not arguing over the meaning or utility of the observation.
Parsimony is a means of reconciling the distribution of observations, and
we need no other tool.
In my limited experience, older observations are usually retained,
perhaps in modified form, in new analyses. When they are not, it is
often because of a misplaced instinct to discard poorly worded
characters or those which cannot be scored for some taxa. The data matrix
record of a group is little more than the accumulated knowledge of the
anatomical particulars of a group, in a codified and pre-sorted format.
Just because the wording of characters, and the interpretation of their
linkages and polarity, are different does change this. The very fact that
the raw data changes so little, although it does compound over time, says
to me that systematists are all on the same page observationally, although
some may be more perceptive or careful than others.
As for concentrating on the few, important characters, well,
that is what cladists do, we just do it after we have allowed an objective
sorting proceedure to process the data and find trees. We let logic
guide the analysis, which does provide a measure of objectivity.
Logically, the more data we have, the stronger our phylogenetic signal
should be. So don't decry characters, go out and find them!
P.S. Dinosaurs dinosaurs dinosaurs!
P.P.S. Why don't we talk about what happened to dinosaur communitites
between the Early and Late Jurassic?