Re: Cladobabble

["Jonathan R. Wagner" <znc14@TTACS.TTU.EDU>]

At 08:49 PM 10/11/00 -0400, philidor11 wrote:
>I broke this paragraph up because the second part contradicts the first, as
>I read them.
        I took liberties with your text as well. Hopefully yours is as
libertine as mine! ;)

>-the analyst selects which characters are meaningful
        Ahah! This is not what I meant. The analyst makes observations of
features which appear to be present in more than one taxon. He does not
judge which is or is not meaningful. Outgroup comparison will suggest which
character state is ancestral and which state(s) is (are) derived. But only
analysis of the most parsimonious tree(s) (and, perhaps, less parsimonious
trees or some consensus of them) will suggest which characters are
"meaningful." this is, of course, assuming that "meaningful" means "helpful
in elucidating phylogeny."

[...]
>and
> --(the tough one to interpret) assuring the significant aspect of the
>character is correctly coded to assure contrast with other samples ('is the
>character polarized correctly').
[and]
>[this assumes that the analyst knows what a significant difference is; for
>example, say how much a bone must differ in thickness to be significantly
>different from a comparable bone of another animal]
        These are different subjects entirely. Your first assertion,
concerning contrast, is worded in a way I feel uncomfortable with.
Polarization is not a contrast issue, it is an issue of utility. The
ancestral state is not useful in elucidating phylogeny, whereas the derived
state is, e.g., if you find 500 taxa with state a and one with state b, the
utility of the character is entirely dependant on which is the derived
state. If it is a, the character is useful, if it is b, it is autapomorphic
(within the dataset), and therefore useless for phylogeny reconstruction
(or, to put it in a snootier way, phylogenetically uninformative).
        As for your latter point, this is, I feel, the only real issue at
the heart of the matter: when is an observation significant? The only anser
I can really give is: when the resulting phylogeny says it is. This is, of
course, unsatisfactory for most people, because you may not see that the
margin 5% increase in bone thickness is a significant difference. If, as I
alluded to in my post, your animals cluster such that this 5% difference or
whatever really represents a great rift in bone morphology in the ingroup, I
don't think anyone will contest the character. Unfortunately, evolution has
a nasty habit of providing us with intermediate, or at least POTENTIALLY
intermediate, steps in any character. So it is very hard (contrary to the
implications of my earlier post) to say that the difference is significant
or not a priori of the analysis. This is why I advocate including all
reasonable characters (reasonable being non-redundant and phylogenetically
informative)
        This is, of course, one of the great things about a cladistic
analysis: if you don't like a character, YOU can rerun the dataset without it.

>-the selection is based on objective standards including
> --whether the form/structure is correctly observed,
        Well, this may or may not be entirely objective (he says, eroding
his own position). Much of our discusison really depends on whether an
observed character is considered a "real" entity or not. I'd say it isn't,
in which case it may be particularly hard to have objective standards for
observation. Fortunately, since characters have to be (or at least, should
be) explained in some small detail, the specifics of the observation are
debatable. Indeed, this is one of the most frequently revised aspects of any
study I have seen.

> --avoidance of characters whose significance is already included
>('manifestation of another character which is already coded'),
>[this assumes a knowledge of the relationship between/among characters]
        Not necessarily. John Merck (currently at the University of
Maryland) has mentioned means of objectivley evaluating character
relationships within a dataset a priori of analysis. I am working on similar
methods, although I can't wait to see his work in print.

>-if a character's form/structure can be observed (the only criterion given
>which applies to importance judgment as opposed to coding judgment), then it
>should not be rejected 'a priori of analysis' of relationship.
        If it cannot be rejected under the criteria I mentioned.        

>Part 2
>-characters should be included regardless of  utility (not sure how you're
>using 'meaning' in this statement) to the animal
        Touche'. I meant it ambiguously, as in, the way "others" might refer
to the "evolutionary meaning" or "importance" of a character, or its
phylogenetic "utility."

>-parsimony alone will determine relationship without regard to any other
>consideration.
>
>Remember that, from Part 1, you have eliminated characters whose
>significance has already been included by the use of other characters.
>Might not the most parsimonious tree be different if you had included the
>other characters?
        Of course it would. However, consider the follwing (excessive) example:
        Character list one:
        Give milk to young: 0) no, 1) yes
        Loss of egg:    0) no, 1) yes
        Number of legs: 0) four, 1) two
        Number of feet: 0) four, 1) two
        Number of appendages not used for ground movement: 0) zero, 2) two
  
        Taxa: Ursus arctos, Homo sapiens, Columba livia

        Guess what, you have discovered a human-bird clade. Never mind that
the last three characters are all manifestations of the same phenomenon
(bipedality), so say I, anyway. This is what I'm talking about... not an
HYPOTHESIS of redundancy among characters, but a non-falsified THEORY of
redundancy.

>Also from Part 1, you asserted knowledge on the part of the analyst of  when
>a difference becomes significant.
        Actually, this is absolutely the opposite of what I meant. In fact,
that was what my post was aimed against. I was not trying to show how
systematists can pick and choose the characters which they want, I was
trying to show how they SHOULDN'T. I think if you reread the post from that
perspective, you may find we agree on more issues than you originally thought.

>Isn't that an a priori decision of
>non-relationship? 
        Actually, the idea is not to judge which characters are not
significant, but to observe features, thereby hypothesizing that they MIGHT
be significant. Once observed, rejecting the potential homology a priori of
analysis should require argumentation, rather than offhand dismissal. Sadly,
no animal comes with a list of characters attached. All we can do is try to
find characters, and use this positive evidence to refine our hypotheses. We
certainly should not expect to have all the characters at any one time,
which is why a phylogenetic hypothesis is necessarily a work-in-progress...
it represents the most logical arrangement of the taxa used based on the
morphological observations made UP TO THAT POINT.

>If I have 2 animals and one has a bone twice as thick as
>another's (my hypothetical example), by the very act of coding the
>distinction, have I not made this a difference which will guide the
>algorithm in working out a connection between the animals?
        You have observed a difference. If you can identify a similar state
in another taxon, you have observed a potential shared, derived charachter.
If you can determine that the shared character state is indeed derived (or
the polarity is equivocal), you have identified a potential synapomorphy.
        I see your problem, and part of this has to do with the nature of
dealing with observations using an equally-weighted parsimony. If
observations are not real, than our use of each one does, theoretically,
unweight the data.  The only way to find the "true" weight of a character is
to find the TRUE tree of life, and map all characters onto it.
Unfortunately, since characters aren't real, there is no set number of
characters for that tree, so there is no set weight for each character. In a
climate of ignorance such as this, we are best off trying to reduce our
observations to those which can be reasonably believed to be independant,
and then apply parsimony.
        So, yes, coding the character gives it weight, but ther is no way
around that. In fact, in a sense, that is EXACTLY the point of the exercise.
What we want to do is avoid unduly over- or under- weighting that character
by excluding observations based on a priori judgement, or including
redundant characters (respectively). This is why I argue that all possible
characters which cannot be reasonably eliminated should be included.
        To look at it another way, the act of observation allows a character
to become significant, more singificant than the observations we have not
made yet. However, this is no different than in any other branch of science,
as far as I can tell. I'm no philosopher of science, but it seems to me that
we cannot allow ourselves to be burdened by all the things we do not know,
nor can we let our awareness of them prevent us from trying to make sense of
what we do know.

        Hope this helps!

        Wagner
--------------------------------------------------------------------------------
     Jonathan R. Wagner, Dept. of Geosciences, TTU, Lubbock, TX 79409-1053
  "Why do I sense we've picked up another pathetic lifeform?" - Obi-Wan Kenobi