[Taxacom] On pattern and structure

Jason Mate jfmate at hotmail.com
Tue May 24 17:39:51 CDT 2011












Yikes, talk about jumping from the pan and falling on the flames!
"Ah, excellent, Jason. Yes, 
phylogenetic is structuralism. The "phylogeny" is the problem. Take a 
molecular phylogeny. It isn't a phylogeny, it is a rather precise 
assessment of present-day relationships of exemplars. The present-day 
relationship of exemplars is not a fundamental structure in nature 
(i.e., a structuralist pattern), it approximates a phylogeny only to the
 extent that it estimates genetic continuity and branching order. 
Speciation is not estimated, and there are artifacts like paraphyly and 
phylogenetic polyphyly. "

In regards to what a phylogeny is, it is by definition a construct that minimises the number of steps (changes) necessary to account for the variation in the group of taxa that you have available to study (extant or otherwise). The phylogeny, any phylogeny for that matter, can only account for the known, e.g. the taxa at hand. What happened during the evolution and has been lost can sometimes be inferred but can never be known. I can think of examples where an acrimonious debate regarding the evolution of certain features raged for decades til the discovery of certain fossils settled the matter(i.e. bird evolution) . Based on the available evidence several scenarios were possible so the addition of the new fossils gave them a peek at the past and settled the matter.
Hence a phylogeny is a rather blunt tool, an idealised representation of the process of evolution of the sample. It does not (usually) show the finer processes at the heart of divergences (tokogenesis), hybridisation/reticulation, etc, although methods are available to detect the latter (more or less). I am not sure how you would go about estimating speciation outside the species that exist or the fossils that you have available.

It is not fundamental because morphological 
cladograms are rejected and all morphological data are simply mapped to 
the molecular cladogram (i.e., no theory is presented on why they don't 
match, although you can read allusions to convergence and incomplete 
lineage sorting). There is no recognition of macroevolution, which is 
fundamental. Patterns of macroevolution as devised from all relevant 
data should be that fundamental pattern on which one can analytically 
map data that are dependent, but we get only microevolution of traits 
mapped on a molecular cladogram, with no allowing for the fact that 
about 63% of extant taxa have extant ancestors. 

These problems are common to all data so I have to disagree with your assertion that morphological data are somehow deemed as second rate (though I acknowledge that there are practitioners who feel that way, their loss). It is true that characters under scrutiny (usually morphological or ecological) are mapped onto phylogenies derived from other data, but his is to avoid circularity. As for macroevolutionary patterns, I am not sure what you mean: of the group at large or maybe alluding to some universal principle (i.e. the relation between parental care and clutch size). Likewise the last statement somehow feels out of place. I am guessing that you are alluding to stem species and the extinction of stem species upon speciation events (sensu Hennig). This is a species concept issue and I don´t feel like the use of phylogenetic methodology makes it a de facto acceptance of a strict cladist species´ concept. 
Let's take the long-suffering polar bear and two molecular brown bear lineages as an example. The evolutionary theory and fundamental pattern is actually an explanation of all data, the polar bear came from one of the brown bear lineages, and Dollo evaluation supports this. The molecular cladogram ((polarbear,brownbear1)brownbear2)... is not a hypothesis, it is a fact. And there is no explanation.
 
A hypothesis is not a fact. What the cladogram tells me is that the polar bear is nested within brown bears. What I read from it is that Polar bears evolved from a lineage of (lets assume we have this info) northern brown bears and not southern brown bears. And from this we could hypothesize several different scenarios as to how they evolved, and hopefully do so in a way that they are testable, if not with the available data then with future study. From the taxonomic/systematic perspective, a strict cladistic (sensu Wheeler) would argue that either the polar bear is just a brown bear or that each bear lineage is a separate species. Cladistics are not at fault though, they follow Hennig and his naive (some would say simplistic) view of speciation. However I think that Hennig understood that speciation is not quite so clean but wanted to come up with an objective species concept. Of course, decades later it is just one more definition and not one of the better ones at that.


I'm reading a fine book by Morris Kline, Mathematics and the Search for Knowledge. In one chapter he points out that mathematics is totally fundamental in physics. In fact, in electromagnetics, we have a number of fundamental formulae that describe and predict electromagnetic phenomena. But we don't have ANY idea of what it is that propagates a force at a distance like gravity, or what constitutes magnetic field lines, or what is vibrating in wave theory, like what moves in a radio signal. Kline asserts that even so, the science based on mathematics is stupendous, and causal, essential, mechanical knowledge, though absent, has not proved important. 
 

Well, okay, and the detailed and precise phylogenies of molecular taxonomy are indeed immensely mathematical, but these patterns are NOT mathematical formulae such as there are in physics. Such phylogenies conflict with morphological phylogenies (explanations are eschewed in structuralism), and internally are inconsistent in evincing paraphyly and phylogenetic polyphyly of what are clearly one taxon (in most cases, IMO). 
 
I don´t see the connection here. The phylogeny (whatever the data) is the product of simple algorithms based on generally accepted axioms (parsimony). Like other combinatorial optimization mathematical problems (i.e. minimum spanning trees) it seeks to connect the samples in a way that is least costly. Similarity to the real phylogeny will be dependent on the information. Alas most organisms and their geno/phenotypes are gone so you just have the most recent editions (and only a few exemplars at that) to work out the path. Conflicts probably have as much to do with lack of data as with problems of homology, homoplasy, etc. But what the phylogeny doesn´t tell you is if your species should be monophyletic or paraphyletic.



You may well infer a macroevolutionary scenario from the polar bear, brown bear molecular analysis, and some may not, but my beef is with the phylogenetic classification that results: namely, lumping the polar bear with the brown bear, or recognizing three taxa at the same rank, two of them cryptic and silly. Taxa without names, or those lost among multiple cryptic taxa, cannot be part of a biodiversity analysis, or be conserved.
 

It may be that problems like the paraphyly of the brown bear, and the portulacca family, will be presented with work-arounds, but there are myriad major groups that have been lumped because recognition requires formal reflection of macroevolution in classification. In my own group of expertise, three families have been lumped into a fourth larger family, Pottiaceae, with no explanation other than that they are nested. There are five huge families also nested but near the base of the Pottiaceae, and phylogeneticists don't have the nerve yet to lump them, but I'm not waiting until they do so to complain about it. If systematists of good will say nothing, now, structuralism will prevail. 


I have no problems recognizing  paraphyletic species or higher groups if it is the best choice available. Obviously polar bears are distinct (i.e. we can recognize them as distinct species based on different species concepts) so the fact that it fails at one specific concept is no reason to make counterintuitive decisions. I am not familiar with Portulacca or Pottiaceae but, bringing up one of the most familiar example of an unnatural groups (reptiles) splitting this one is not unreasonable. Similarly many monospecific genera have been erected in the past and then found to be deeply bested within an otherwise monotonous group. Again, I take no issue with lumping, since the recognition to supraspecific taxa is often a subjective criterion to make biodiversity more manegeable. To me practicality and common sense should be guide in using phylogenies to create classifications. That said I would still not accept a polyphyletic group.

Best

Jason



 		 	   		  


More information about the Taxacom mailing list