[Taxacom] Mona Lisa Smile

pierre deleporte pierre.deleporte at univ-rennes1.fr
Mon Jul 31 07:09:29 CDT 2006

A 21:19 30/07/2006 -0500, Thomas Lammers wrote:
>----- Original Message -----
>From: pierre deleporte <pierre.deleporte at univ-rennes1.fr>
> >> And if you could specify at little bit more which kind of "overal 
> similarity"<<
>I want a classification that takes into account ALL aspects of a species 
>biology: morphology, chemistry, ecology, etc.  If the phrase "overall 
>similarity" bothers some, I'll rephrase it as, "the most things in common."

Thanks for precision, but my hint was that there is an illimited number of 
ways to classify phenetically on the basis of the same data matrix, and I 
expected your specifying which way and why... Thus I'm left in complete 
ignorance of exactly how you are classing and why this specific way (note 
that I have nothing against deliberately accepting some arbitrariness, the 
real sin consists in feigning to ignore it).

Now, we were talking of supra-specific levels. The problem is then your 
first reference to phylogeny, seemingly corrected in favor of overall 
similarity. It is demonstrated that overall similarity will not provide you 
with a phylogeny, except when evolutionary divergence is regular 
("clocklike" evolutionary process). Hence I suggest that you present your 
approach as phenetic, and dispense completely with phylogeny, which can 
only be an occasional by-product of a phenetic clustering. You have really 
no need to start with a phylogeny if you need overall similarity.

>In supraspecific classifications, I think evolutionary history must be the 
>guiding basis.

This is where I have difficulties to follow you logic: if you propose this 
phylogenetic approach, it is at odds with phenetic "overall similarity" 
classification. You seem to be trying to implement two incompatible logics 
without clearly deciding of a priority rule.

>As I said, more often than not, this should result in classifications in 
>which the taxa grouped together have the most tings in common.  I will 
>accept only monophyletic groups -- in the REAL original meaning of 
>"monophyletic" -- i.e., all members of the group share a common 
>ancestor.  When some descendents of a common ancestor have diverged quite 
>a lot, I kick them out of the group.  What is left behind is still 
>monophyletic in my book.

OK this is what is generally called "paraphyletic", but if you call 
paraphyly monophyly, then you should specify it for intelligibility and you 
could also call your complete crown groups "holophyletic" (and Ken would be 

>    Anyone who banishes paraphyletic groups is ignoring how evolution 
> operates.

Certainly not, merely because there is no logical connection between 
knowledge of evolutionary processes and decisions about classificatory 
rules (see also Richard Pyle's further comments about what is biologically 
testable versus nomenclatural conventions: mixing the two points is 
obviously a lasting source of great confusion in systematics). Note that it 
is neither sure that anybody who accepts the possibility of paraphyletic 
groups in classification is aware of how evolution "operates" at all. But 
two biologists can obviously be in complete agreement about evolutionary 
processes and phylogenetic topology in a group, and use different 
classifications (e.g. with, or without, paraphyletic groups, according to 
their practical needs).

> > But my precise point is: for which purpose do you consider your system 
> to be "the best classification"?<
>For ALL purposes, biological and otherwise.

Sorry but this is not an answer, just a plain statement. Why should I trust 
you, that such a mixture of phenetic and phylogeny (with still no clear 
criteria for combining the two requirements) is best for all purposes? You 
should at least provide some examples of use in some specific contexts.
I am suggesting exactly the reverse approach in fact: specify one or 
several contexts of use for systematics, establish the specification (list 
of desired properties) for an optimal system, and then devise this system 
(or at the very least check whether your own system works better than some 
others in this context).
A mere statement "it is the best" has strictly no convincing force, except 
contributing to make hopeless pseudo-debates last a little bit longer, 
because it calls for and legitimates in advance the obvious answer from 
another eclecticist with universal pretentions: "no, it's my system which 
is the best for all purposes".

>   If "best" offends, substitute "most useful."  The only thing that would 
> carry more information than such a classification is a phylogeny.

Hence the later should be "the best", because you need a classification 
carrying "more information"? But as a rule overall similarity will always 
be as 'informative' or even more 'informative' on more features than mono- 
or paraphyletic grouping, precisely because it is "overall" and not limited 
to the series of original synapomorphies plus uncontrolled proportions of 
shared symplesiomorphies and further evolved features.

>  But phylogenies make clumsy classifications.

I suggest you define "clumsy" in biologically intelligible terms, and in 
some context of relevance, unless this point of yours is not debatable.

> >>Why not everybody his own optimal system rather than an inevitably 
> sub-optimal, if not far from optimal, compromise?<<
>So that we have a common touchstone for communication.  So that we do not 
>recreate the Tower of Babel.

One can have both eclectic or monophyletic classifications written in 
Esperanto, hence universal communicability is guaranteed. Everybody can 
also hopefully agree on a specific self-consistent system for common use in 
a specified context, hence biological universality is preserved. 
Universality means universal agreement inside every commonly recognised 
context of relevance, not necessarily throughout all imaginable contexts.
In my view classifications are not properly "general language" in science, 
they are specific tools which can be optimal only in specific contexts. 
Hence I consider that you are sacrificing optimality (of different 
classifications for different contexts of use) in favor of "common 
language", which is simply not the point.
A common data matrix for terminal taxa (I won't open the latter can of 
worms) plus deposited types for unambiguous biological reference and 
nomenclatural rules against ambiguous naming (codes of nomenclature) make 
sufficient "common language" and reference for all uses, whatever the 
higher classification you will adopt for convenience.

Now isn't such a "classificatory Babel tower" in front our eyes after all, 
when no classificatory logic is imposed by the codes, hence everybody is 
perfectly free to class his own personal way?
Are you planning to impose an "Eclecticocode" (your personal vision of an 
optimal combination of phenetics and phylogenetics, the 'Lammers system') 
against some "Phylocode" and also against competing eclectic systems 
(Kinman's...)? Because Ken makes no explicit use of overall similarity, 
rather some intuitive notions like what a true bird is all about - which 
tastes very much of implicit "natural classification", by the way.
Or will you propose your system alongside with Ken's one, contributing 
yourself to what you call a sytematic Babel tower? Or suggest an 
international jury? (The only way for imposing arbitrary decisions I guess.)

A unique classification cannot be optimal, not even convenient, for all 
purposes, because of mere logics: by definition, "optimal" and "convenient" 
as necessarily context-dependent. Hence I have great doubts about the 
interest of imposing a unique system. A strict "overall similarity" 
(phenetic) system of some kind could imaginably be optimal in possible 
contexts where precisely optimal overall similaity is required, but when 
you renounce to applying this criterion uniformly and combine it with 
phylogenetic considerations, I'm lost.

There is nothing like a personal charge in my comments, I think that 
probably most of contemporaneous systematicians don't bother very much 
about why they class this or that way. They were simply not trained to ask 
such preliminary questions, beyond the classic trilogy "eclectic, phenetic 
or cladistic" presented as competing rather than possibly complementary, 
not to mention doubtful notions of "true scientificity" awkwardly applied 
to practical matters of convenient classification. The illusory belief in a 
self-evident "natural classification" is seemingly still a heavy implicit 
cultural burden, partly but deeply rooted in the notion of the query for a 
unique universal system, while inquiring extensively about one's needs 
before adopting a system or another seems still to be a very exceptional 
practice. Hence my own pointing at Tom's or Ken's implicit motivations, or 
Christopher's supposed needs, should be viewed as illustrative examples.


>Tom Lammers

Pierre Deleporte
CNRS UMR 6552 - Station Biologique de Paimpont
F-35380 Paimpont   FRANCE
Téléphone : 02 99 61 81 63
Télécopie : 02 99 61 81 88

Taxacom mailing list
Taxacom at mailman.nhm.ku.edu

More information about the Taxacom mailing list