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

Re: More on classification.



 > Last week Stan Friesen  <swf@tdat.ElSegundoCA.NCR.COM> presented a
 dinosaur
 > classification and then an explanation of its rationale
 > ..
 > >So, I presented my alternative classification.  It is
 > >more efficient as a filing system, since it has fewer levels,
 > >and on average 3-5 subgroups for each taxon, instead of the
 > >2-3 of the cladistic classification.  (Cognitively, 3-5
 > >subgroups is more 'natural' for the human mind than two).
 > ..
 >
 >     If the only reason that we classified taxa was to have good ways
 to
 > remember them then we could use all kinds of criteria,

Quite, if memory was the *only* criterion, that would be true.

But if it is merely one of many criteria of a good classification,
then that is not true.

Another criterion of a good classification, in biology, is that
the organisms be related.  This is captured in the requirement
that taxa be "monophyletic" (in the *original* sense, not in the
revised cladistic sense).  Thanks to the cladistic redefinition
of the term, there is now no term that refers to what "monophyletic"
used to.  That is, paraphyletic taxa are monophyletic in the original
sense of the word, and there is no term meaning "cladistically mono-
phyletic or paraphyletic".

The following is an outline of the criteria mentioned in
"Principles of Systematic Zoology, 2nd ed."

        - A classification is an index to stored information
        - A good classification has heuristic properties
          [where heuristic means "allows one to make predictions"]
        - Classifications permit the making of generalizations
        - A classification has explanatory powers

A cladistic classification really only meets the last criterion.

 > But I felt  moved to write because the tone of Stan's
 > note made it sound as if the issue between cladistic/phylogenetic
 > classification and his was one of convenience.

Well, in part, also information content (broad sense), and the
making of useful physiological and anatomical generalizations
[that is heuristic value].

I maintain that cladism is weak in all of these areas.
 >
 > >[a cladistic classification] results in unbalanced classifications,
 > >with a few large, morphologically heterogenous groups, and many
 small,
 > >slightly differentiated groups.
 >
 >       But the important insight that is being missed here is that
 since
 > the middle of the last century we have realized that all organisms
 > are part
 > of one tree of life,

No, I am *not* missing that.  ALL biologists use that fact in
their classifications.  It is not, in any way, shape or form,
a difference between cladistic and evolutionary classifications.

The difference is that cladists allow only *branch* points to
matter, evolutionary taxonomists allow the "length" of a branch
to matter - that is the amount of difference along it.

That is why I concentrated on what I did - it is the core of the
*difference* between the two philosophies.  The aspect you are
talking about is universally agreed upon, and thus does not
constitute a difference.

[Note, while details of the presented classification were mine,
the basic structure came from the first of the references I cited,
the "Mesozoic Meandering" one].

 >  this tree is a series of bifurcations

Change that to polyfurcations.

If you take the Ernst Mayr model of speciation seriously,
bifurcations are not at all the only mode of branching.
The appearance of bifurcations is, in part, due to the
very methodology of cladism, which tends to force resolution
of multi-branches into bifurcations.
 >
 >     There is a further issue of whether we should use new names when
 a
 > particular linneage has changed a lot (e.g. birds).

Actually, that is the *only* issue.

That is the only place that evolutionary taxonomists and cladists
differ.

Now, for a discussion, in mathematical terms, of why such a
practice is desirable, and why it leads to classifications with
better *predictive* [heuristic] value (with regard to non-taxonomic
characters), I suggest the following article:

        Carpenter, Kent E., 1993, Optimal cladistic and quantitative
        evolutionary classificationms as illustrated by Fusilier
        Fishes (Teleosei: Caesionidae).
        Systematic Biology 42(2):142-154.


In fact a very minor change to taxonomic practice can derive an
evolutianary classification from a cladogram.  that is, at any
given taxonomic level, split the cladogram into contiguous sections
at those branches with the largest weighted character count.

One simple method for doing this can be found in tha appendix to
chapter 9 of "Principles of Systematic Zoology, 2nd Ed." by Ernst
Mayr and Peter Ashlock.  [This method was developed by Dr. Ashlock
and one of his grad students, and is similar to the one sued by
Dr. Carpenter in his comparison in the above mentioned article].

 > Sober, E.: 1988, _Reconstructing the Past.  Parsimony, Evolution,
 and
 > Inference_,MIT Press, Campbridge.

Uh-huh, no mention of maximum likelyhood estimates, just parsimomy.
[I do not trust parsimony, since mosaic evolution is so common].
 >
 > Ridley, M.:_Evolution and Classification.  The Reformation of
 Cladism_,
 > Longman, London.

Are you sure you want to recommend a book on reformed cladism?
That is merely a thinly disguised version of phenetic taxonomy,
and is not, philosophically, really cladism.
[Unless this book is really not on what is commonly called
reformed cladism].
 >
 > Gauthier, J.:1986,'Saurischian Monophyly and the Origin of Birds',in
 K.
 > Padian (ed.), _The Origin of Birds and the Evolution of Flight_,
 California
 > Academy of Sciences, San Francisco, pp 1-55.

I seem to remember reading an article with a title like this,
was a version of this published somewhere else?


swf@elsegundoca.ncr.com         sarima@netcom.com

The peace of God be with you.