pierre deleporte pierre.deleporte at UNIV-RENNES1.FR
Mon Oct 6 14:31:47 CDT 2003

Thanks to Richard Zander, I have not to devise an example...

Just a comment : this is why, in my view, phylogenetic classification 
should be clearly disconnected from classifications based on shared 
characters and character states. As suggested recently by Don Colless, a 
phylogenetic classification merely consists in representing the 
phylogenetic tree. Period. Drawing the tree could be sufficient. Turning 
the tree into codes or words is possible.

Much confusion comes from the attempt to conciliate a phylogeny-based 
classification and a character-based classification. Synapomorphy clearly 
doesn't always work as a universally shared feature in phylogenetic 
classes. Not to mention alternative ways (other than parsimony 
optimization) to infer a phylogeny from a data set.

The one and only property shared by all members of any phylogenetic class 
(=monophyletic group) is: to be descendant from an exclusive common ancestor.
This is a property directly inferred from an already accepted phylogeny 
(historical explanation of life), this is not a material character state 
carried by specimens. Trying to conciliate both logics an exact way is 
doomed to failure.

In my humble opinion (at the present time).


A 16:34 04/10/2003 -0500, Richard Zander wrote:
>Well, heck, you're right. Taxa that share no advanced character states in
>common can be interpreted as part of a monophyletic clade: Try this with
>begin data;
>         dimensions ntax=12 nchar=18;
>         format symbols="01";
>         matrix
>0 000000000000000000
>A 111111110000000000
>B 011111111000000000
>C 001111111100000000
>D 000111111110000000
>E 000011111111000000
>F 000001111111100000
>G 000000111111110000
>H 000000011111111000
>I 000000001111111100
>J 000000000111111110
>K 000000000011111111
>         ;
>outgroup 1;
>set maxtrees = 200000;
>boots nreps = 2000;
>The bootstrap is interesting in that taxa in the middle of the lineage (e
>and f) are least well supported.
>Richard H. Zander
>-----Original Message-----
>From: Jan De Laet [mailto:jdelaet at NATUURWETENSCHAPPEN.BE]
>Sent: Friday, October 03, 2003 3:13 AM
>Subject: Re: [TAXACOM] necessary?
>Nico, you can rest your case, as far as I am concerned (well, apart
>from your botanical example, as Paul van Rijckevorsel pointed out).
>I would just like to rest it a bit more: the theory of cladism
>(parsimony analysis) does NOT require adherence to essentialism. As
>you correctly note, "unique synapomorphy and perfect congruence are
>not "necessary"". This is true in practice as well as in theory
>(and one does not need to be a Sherlock to figure that out :-)).
>I think it is also worth pointing out explicitly that your note
>shows that parsimony analysis has nothing to do with monothetic
>classification. Being monophyletic and being monothetic are
>simply two different things, and the theory of parsimony analysis
>does NOT rely on the the methodological assumption that 'monophyly =
>monothetic', as Richard Zander seemed to imply the other day.
>Of course, if one happens to find unique synapomorphies for clades,
>that's nice, not in the least for practical reasons. But parsimony
>analysis does not assume that such things exist, not in theory and
>not in practice. And even if, in a particular analysis, unique
>synapomorphies are found, they don't have to be interpreted in an
>essentialistic way. Why insist on that interpretation? Looks like
>a strawman to me, or some severe misunderstanding of current theory.
>Jan De Laet
>Freshwater Biology
>Royal Belgian Institute of Natural Sciences
>Vautierstraat 29
>"Nico M. Franz" <nmf2 at CORNELL.EDU> wrote:
> > If tomorrow an orchid species would lose its triploid endosperm or an ant
> > evolved another pair of legs, we would still be able to refer to these new
> > taxa as angiosperms and insects, by virtue of the remaining associated
> > (congruently nested) properties they possess. In practice this happens all
> > the time. We call snakes tetrapods in reference to their shared
> > evolutionary history and attributes (perhaps even the underlying genetic
> > basis for legs which they should still have) - although they are "legless"
> > phenotypically.
> >
> > Again I'd commend the article by Boyd. Higher taxa can be individuated in
> > reference to homeostatically maintained sets of properties. Often these
> > properties have a causally sustained yet imperfect association in space
> > time. The secondary loss of one among many of them is not necessarily an
> > obstacle to reference. In this sense evolutionary naming may differ a bit
> > from the way in which we name gold.
> >
> > In practice unique synapomorphy and perfect congruence are not
> > as I think the snake example shows. Homoplasious characters play an
> > enormous role in cladistic reference. If the theory of cladistics requires
> > the adherence to essentialism, then it seems at least that actual practice
> > hasn't paid much attention to that theory (can I now rest my case?)
> >
> > Regards,
> >
> > Nico Franz
> >
> > At 12:09 PM 10/3/2003 +1000, you wrote:
> > >The notion that classical taxonomy is essentialistic, is of course (as
> > >pointed out by Nico Franz) nonsense. Its polythetic nature was examined
> > >back in the 60's, by inter alia, David Hull (in what I seem to remember
> > >was his PhD thesis!). The amusing fact is that cladism is absolutely
> > >essentialistic. Its taxon names are rigidly defined by a set of necessary
> > >and sufficient conditions: the posession of what are BELIEVED to be
> > >apomorphies (by, of course, a set of self-appointed Sherlock Holmes').
> > >
> > >Div of Entomology, CSIRO,
> > >GPO Box 1700,
> > >Canberra. 2601.
> > >Email: don.colless at
> > >Tuz li munz est miens envirun
> >

Pierre Deleporte
CNRS UMR 6552 - Station Biologique de Paimpont
F-35380 Paimpont   FRANCE
Téléphone : 02 99 61 81 66
Télécopie : 02 99 61 81 88

More information about the Taxacom mailing list