[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

[no subject]

David Marjanovic wrote:

<There's nothing wrong with using proportional characters in principle, and I 
don't see why you call them "horribly prone to convergence", or what you mean 
by "tend to group into suites of characters".>

  As I wrote the above quote-marked statements, I feel a defense is in order.

  On the mere basis of a character matrix, any character will do, and it 
doesn't matter what it's a character of. If you're doing analytical metrics, 
it's only relevance to reality is that it is directly identifiable in the 
fossils. No theoretical hypothesis preceding the nature of the character states 
really need be in place. Relating the character to a phylogeny does not even 
impact this selection that much, as it assumes _a priori_ for there to be a 
relevance (this should alwaysd be done after the analysis anyway). I have 
argued in the past that on the flat basis of an ecologically convergent 
feature, characters should be somehow de-weighted, and that ecologically 
correlated feature cannot be determined from a phylogenetic analysis. This 
divorces a set of characters, many of which are proportional, from phylogenetic 
analysis _a priori_. Re-including them after the analysis to test their 
relevance has, to my knowledge, not been tested, although hypothetically it is 
understood that excluding them allows the remaining character signal to be 
read. Convergent characters introduce noise, as do proportional characters (as 
they tend to also be convergent), and one would note that many of these are 
slowly being weeded from analyses or reformulated to not depend on length vs 
length-type models. If you think that's a flaw, that's fine: I certainly do not 
think that exclusion of all "features" is reflective of reality in phylogeny, 
but it obscures the signal that "being big" or "being aquatic" produce in a 
matrix's output. One of the biggest issues of characters being weeded out in 
fact deals with the "robust" marker (bigger animals tend to be more robust than 
smaller ones, but this is a proportionate continuum that is never defined in 
the taxa, and can involve cortical thickness, diameter to length, breadth to 
length, diameter to breadth, the development of a "scar" into a "crest," or a 
"crest" into a prominent "ridge," etc.)

  A further difficulty lies in the "suite" concept of features. It is almost 
certainly impossible to determine the envelope of a character suite, as any 
skeletal system is never isolated from another, and any change in one bone 
affects others (or character relating to others, such as proportions). Some 
taxa may even have the same set of features, but have an additional feature 
that divorces a feature from the suite, or marries two suites into one, and 
this tends to occur in pnuematics as systems are integrated or defined from one 
another. Laminar systems in sauropod vertebrae are almost certainly integrated 
with the camarate system in their vertebrae, as are the muscle or apophysis 
positions, yet these are all different sets of suites that are nonetheless 
wholly integrated with one another and dependant on each other: High surface 
diverticular expression should result in high laminar expression, which follows 
from high internal camarate morphology of the vertebra, but also results in 
bone reorginization to deal with bone stresses, and results in apophyseal and 
attachment site movement (and is therefore also proportional). The question of 
diverticular convergence (presence of distinct laminae, camarae, etc.) is done 
through secondary grouping of features in a matrix, while two taxa with wholly 
identical laminar structures can derive from two different lineages without 
said laminar structures. The "suite" concept is both useful and dangerous, and 
should, regardly of Weins or Marjanovic and Laurin, be given a broad berth for 
analysis whenever it is included.

  And no, I have no specific objections to either Weins OR Marjanovic and 
Laurin analyses. They are both useful, in context. The caveat is simply that 
the very nature of convergence, while being potentially phylogenetically 
informative, is still convergent and the cladistic analysis cannot determine 
the difference without _a priori_ input on the part of the observer, which 
tempers the character analysis after the phylogenies are produced. Tree 
selection, by its nature, looks for the shortest tree with the least reversal, 
and uses the products to temper the theory that such is close or ideal in 
studying the concept of said reversals, and is thus self-cannibalistic in 
regards to the character input and the interpretation of the character suites.

Thank you, and cheers,

  Jaime A. Headden

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

"Human beings, who are almost unique in having the ability to learn
from the experience of others, are also remarkable for their apparent
disinclination to do so." --- Douglas Adams (Last Chance to See)

"Ever since man first left his cave and met a stranger with a
different language and a new way of looking at things, the human race
has had a dream: to kill him, so we don't have to learn his language or
his new way of looking at things." --- Zapp Brannigan (Beast With a Billion 

Your E-mail and More On-the-Go. Get Windows Live Hotmail Free.