[CPN] proposed revisions of Note 9.3.1

[David Marjanovic]

...Definitely food for thought, both the proposal and the manuscript Kevin dQ graciously attached.

Three parts: musings, concrete points about the current proposal, and a little footnote.

==================

Musings:

I have always imagined phylogenetic trees as lineage trees that turn out to be abstract, low-resolution representations of relationship trees -- family trees, genealogical trees -- when put under a microscope.

The manuscript works best under Hennig's species concept, where "species" is simply a synonym for "internode"/"branch", so that "speciation" is simply a synonym for "cladogenesis". Most of it also works under others -- anagenesis and Gouldian budding are mentioned in a parenthesis. However, much of it breaks down under concepts where species can be larger than an internode or contain parts of two or more internodes, so that not every cladogenesis is a speciation. You might think that all such species concepts have long been extinct, but they're alive and kicking. It's easy to find ancestor-worshipping paleoanthropologists who still hold that kind of species concept: here's a massively precladistic phylogenetic tree http://www.globalchange.umich.edu/gctext/Inquiries/Inquiries_by_Unit/Unit_5_files/image017.jpg that shows *Australopithecus afarensis* giving rise to *Paranthropus aethiopicus*, _staying_ *Australopithecus afarensis*, and then giving rise to *A. africanus* (possibly by Gouldian budding), which then gives rise to *Homo rudolfensis*, stays *A. africanus*, then gives rise to *H. habilis* (and yes, this makes the genus *Homo* POLYPHYLETIC), _still_ stays *A. africanus*, and _then_ goes extinct hundreds of thousands of years later. If _such_ a species is allowed to be, in its entirety, an ancestor in a phylogenetic definition, we can get all sorts of unintended consequences. Therefore, I had always thought that species were implicitly outlawed as ancestors, and that "ancestor" referred instead to... Mike Keesey, I'm sorry, I forgot what has become of your "ancestral sets". :-]

Maybe we should simply go ahead and explicitly define "ancestor" as some such "last part of an internode" set that does not necessarily correspond to an entire species under any species concept. I have no idea, however, if that would make the PhyloCode (ICPN) more difficult to accept for a significant number of people.

Anyway, the only problem I can see with "minimum clade definition" and "maximum clade definition" is that those might be interpreted as different ("minimal and maximal") definitions of the same name (or, worse, the same _clade_). How many biologists are there who know graph theory better than phylogenetic nomenclature and therefore interpret, as Martin et al. have done, "node-based" as "relationship tree" and "branch-based" as "lineage tree"? This is an honest question, I have no idea.

"Maximum crown clade definition" might be even more confusing, in that such a definition is a minimum clade definition whose specifiers are pointed at by a maximum clade definition. I'm not sure what to do about that, though. Maybe Note 9.3.1 should be reorganized altogether in a way that avoids this potential confusion. I like, unsurprisingly, the way I explained it in the Wikipedia article on phylogenetic nomenclature; here's a link to a version I trust to say what I mean http://en.wikipedia.org/w/index.php?title=Phylogenetic_nomenclature&oldid=514176633 . This explanation also assumes that "ancestor" cannot refer to an entire internode. The article has since been modified in ways I have yet to understand, for instance ancestor-based definitions ("ancestor per se definitions" in Kevin's manuscript, with a confusing lack of hyphens) aren't mentioned at all anymore in the current version of that article, despite their importance to the rest of the explanation.

==================

Concrete points about the current proposal:

I am particularly happy about the replacement of "most/least inclusive" by "largest/smallest". The former are unambigous, but sound abstract enough that -- for a long time -- they managed to confuse me anyway.

In the proposal to change the definition of "apomorphy-based clade", replace "synapomorphy" by "autapomorphy" (twice). Hennig liked inventing terminology, and he wanted to express every possible concept in a single word made from Greek components; therefore _one_ clade has autapomorphies (auto- = "self") while _two_ sister-groups (or more in case of a hard polytomy) have synapomorphies (syn- = "together"); the synapomorphies of two sister-groups are automatically autapomorphies of the smallest clade they form together, which makes the terms redundant in many cases, but still, there they are, and one clade can't have _syn_apomorphies together with just itself. -- The use of "apomorphy" in that section is correct; that term just means "derived character state" without saying derived relative to what.

By using "and" in strategic places, the proposal to change the last point of Article 2.2 implies that total clades must contain entire species (even if they contain other organisms in addition). In turn, this implies that there cannot be clades within a species. This is correct under Hennig's species concept, but not under whatever concepts the ancestor worshippers think they use. Simply use "or" like in the proposal to change the preceding point (the one about crown clades).

==================

The ancestor-worshipping precladistic paleoanthropologists, which is most of them ("the closer you get to humans, the worse the science gets"), were recently roundly mocked by the tyrannosaur phylogeneticist Thomas R. Holtz, Jr., on Twitter and Facebook: "Australopithecus sediba a mosaic? If only we had a method to sort out phylogenetic relationships..."