Underlying basis of classification (Was: Re Dinobirds)

[Philidor11@aol.com]

In a message dated 7/13/99 12:58:30 AM Eastern Daylight Time, 
qilongia@yahoo.com writes:

<< The naturality of the names themselves is at question, not their 
monophyly; that is, one should perceive that _all_ these names are 
paraphyletic, including Hominidae, Australopithecinae, Lambeosauridae, 
Tetrapoda, Amniota, Canidae, Sarcopterygii, Archaeobacteria, etc., because 
everything stemmed from something else. Taxonomy applies names to better 
differentiate this all to some reasonable order, as far as I see it. >>

The 'reasonable order' you're talking about is implicitly descriptive because 
the animals are ordered by being alike in some way.  You're not specifying 
that the order must be an evolutionary connection, which is the basis for 
cladistics.  Something about this evolutionary system of classification 
bothers me:
Trying to lay this out simply, certain characters are chosen to connect a set 
of animals.  Under cladistics, this relationship is explicitly stated to be 
an evolutionary relationship; that because these animals share these 
characters they therefore share a single common ancestor.  Animals are 
considered closer to (basal) or farther from (derived) this ancestor based on 
the degree to which they have these characters.  The basal connection can be 
quite small; there are basal ornithiscians which do not have the diagnostic 
ornithiscian hip.
The problem for me is the fact that the ancestor is hypothetical.  Without an 
evolutionary progression from documented source to descendant I can't 
distinguish between the diagnostic character and the definition of the clade. 
 Essentially, parsimony says it makes a simpler story if one evolutionary 
progression is postulated rather than any other progression which can be 
hypothesized.  Once that progression has been asserted, though, what evidence 
is relevant to see if it is consistent with known fact?  How much evidence is 
necessary to say that the hypothesis has been supported or even proven?
>From the fact that stratocladistics gets a different name from cladistics it 
seems that including the time when the animal lived is a controversial part 
of hypothesis testing.  Only the characters, it seems,  are essential to the 
analysis, and though character significance has been referred to as 'robust', 
I do wonder about how temporary decorative differences within a group (which 
could be lost in subsequent generations) are distinguished from necessarily 
ongoing functional (evolution being based on functional advantage) 
differences and how a continuing new evolutionary direction can be 
distinguished from simple coincidence in all its various forms (from 
convergence to the sort of repetition which causes certain lizards to become 
legless frequently; might not, for example, a certain group of dinos have a 
tendency to evolve feathers from dinofuzz repeatedly and independantly?).  In 
other words, though a character can be distinctive, how do you know that the 
character(s) is so significant that it indicates a separate evolutionary 
progression from an ancestor to the linked descendants?
When a system of classification moves from a descriptive basis to an 
evolutionary basis, it seems that the evidence of evolutionary progression 
involving these animals (first conclusion) and not those animals (second 
conclusion) should be reasonably close to incontrovertible.  With my limited 
knowledge, I'm surprised at the apparently self-referential standard (is it 
logical within a certain methodology?) being applied.  Can someone help me 
see the strength of the cladistic hypothesis creation and proof process more 
clearly?