The nature of cladistics [...]

pierre deleporte pierre.deleporte at UNIV-RENNES1.FR
Fri Nov 19 17:50:05 CST 2004

Thanks for specifying what you have in mind.
My tentative comment in the text below.

A 06:53 19/11/2004 -0800, Curtis Clark wrote:
>on 2004-11-19 06:21 pierre deleporte wrote:
>>Can there be
>>"processes" affecting a phylogeny?
>Differences in speciation and extinction rates.

These are in fact populational processes (split, extinction).
Maybe the possible "phylogenetic apprehension" of them is just the sum of 
such events in the history of this or that lineage, hence allowing some 
notion of "rate" during a given preriod of time. But while the calculus of 
rates is "phylogenetic" by construction, the processes are populational.

>Differences in inherited
>capacity for morphological change (the inverse of which is sometimes
>called "developmental canalization").

Here we are dealing with genetic constraints on mutations and development 
(genetic stability and developmental "canalization"). Hence processes at 
the genomic level and developmental level. Once more we can check for 
constraints and mutational events in different lineages (phylogenetic 
calculus, versus genomic-developmental processes).

>Differences in the inherited
>features that influence niche breadth, generation time,

Same as above: processes are genetic stability (mutation rates) and 
development. When computed phylogenetically, lasting genetic stability can 
be called "phylogenetic inertia" (differential genetic stability between 
clades). The "real process" is still genomic (mutation), individual 
(development) and populational (selection, drift).

>and all those other things that shape the fates of lineages.

Lineages being not consistent things, what we call their "fate" appears to 
be the sum of events affecting the development of the succession of real 
individuals and real populations in these lineages.

>  And then the totally random things that also affect individuals and 
> populations.

You can also call them "phylogenetic randomness" (versus phylogenetic 
inertia and canalization) when computed at the level of compared 
phylogenetic lineages rather than single individuals and populations.

Note that I'm not contesting the interest of comparing lineages, 
particularly because it's the only way to try and decifer possible general 
laws at the macroevolutionary level.

>Perhaps it's a human construct to say that pterosaurs and glossopterids
>are gone, but turtles and ginkgoes remain.

It's not because it is a human historical construct thait it is not a 
perfectly relevant one, a very likely reconstruction of what really 
happened. I'm deeply convinced that historical science demonstrated that 
pterausaurs existed, and exist no more. But they existed as individuals and 
populations. The point in debate is that this does not make "reptiles 
through the ages" a real thing. It's a set of real things materially 
disconnected in space and time, and clustered in our brains given some 
classificatory criterion (historical property of exclusive common descent). 
This does not impede that similar things share similar properties and can 
undergo similar changes (or lack of) during their time of existence as real 
things. Interestingly, the property of being winged may have contributed to 
some common fate for different pterausaurian individuals and populations, 
but not the fact of  "belonging to the same clade" in itself (this is a 
classificatory notion).

>  And there are people that
>think evolution only happens at a single level (although they can't seem
>to agree whether that level is the gene, the individual, or the

Well, the three of them (plus the cell, organ...). These things undergo 
evolution (change), because they are real, material, biological systems 
when reduced to their contemporaneous living and interacting members. 
Lineages are not. The 'clade birds' does not change (does it?).

>  But if we discount processes affecting clades, we shall
>surely never see them.

In my view we can usefully compare history in different clades (they are in 
fact historical concepts) and check for possible correlation with the 
presence of this or that inherited character in their individual, 
populational... components (wings in pterausaurs...). And very likely you 
and me perform comparative phylogenetic analysis the same way. I simply 
suggest than we can do that  without reifying clades and lineages and 
attributing them processes "at their level", when in fact we mean that some 
common features may cause the same effects. These can be inherited 
features, and then some people will call them "phylogenetic", but 
convergence can do the trict the same way, can't it? If all big flying 
critters undergo extinction in some environmental context, does it matter 
that you "are a pterausaur" at all? Extinction is an individual and 
populational, ecological process of natural selection, not a "phylogenetic" 
one properly. Being winged is a real, biologically meaningful feature 
carried by real individuals. Belonging to a same clade is not.
Concepts like "monophyletic lineage" undergo no processes. I apologize for 
this, but...

Maybe these considerations have no or little bearing on the practice of 
comparative biology with phylogenetic reference (I should chew on this a 
little bit more...). But when it comes to conflicts about the so-called 
"naturalness" of classifications, such confusion (reification of 
classificatory concepts like holo / paraphyly) may become a real burden on 
the debate.


Interestingly, Ken Kinman just posted that we don't invent either holo or 
paraphyletic taxa, when I suggest that we invent them both. The common 
point is that we don't view more "naturalness" in either of them. But 
trying to present them as "real things" is a slippery slope which I fear 
won't help the debate.

Curtis Clark's answer escapes the flawed argument of "naturalness" or 
"reality" of taxa, and he seems to question the criteria for group 
delineation in terms of "significant gaps" ("value" of classifying, and 
"importance" of differences versus similarities).
But without some explicit notion of what we want to do with our 
classification, from which we could draw an explicit specification for the 
classification (i.e. a list of desired qualities), I hardly understand how 
the "importance" of delineation criteria can be debated at all.

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