This is a kind of scattered post on a few things that have caught my eye, while I am avoiding boring work.
Paeloblog reports that a paper in Nature has done a phylogeny on continuous rather than discrete characters, using morphometric criteria to do a hominin phylogeny. This is not the first such attempt to use continuous characters in cladistics, and I would be interested if those who understand this topic comment on this attempt. It seems to me that the main difference between discrete and continuous data would be that the continua are an ordered set of otherwise discrete data points, so there ought to be no problem, logically, but let me know how I have misunderstood (except Envall, of course).
A PLoS Biology paper has also attempted to do functional classification with phylogeny. This seems to me much more dangerous, for a simple reason - functionality implies selection pressure for the character, and hence homoplasy will muddy the phylogeny. But if, on the other hand, one uses a non-functional phylogeny to highlight those things that are functional, there may be much light shed on it. As the authors are trying to rectify the incorrect phylogenies and functional ascriptions to protein molecules, they may be doing something very good here, but I lack time enough to read it carefully.
Here's a paper arguing that African Americans are not noticeably more religious than their white equivalents, but that urban AAs and urban whites are inversely more and less religious in the north and south of the USA. It suggests that religiosity is (wait for it) socially determined. Which we all suspected anyway, but hey...
A letter to Nature points out that south of the US border, Mexican schools teach evolution just fine, and that creationism hasn't made much impact in what is, after all, a majorly Catholic, but secular, country. As I know a few researchers based in Mexico, I am not surprised. Another letter points out that DNA-based classification is founded not on phylogenetics, but on similarity metrics.
And here's a paper that shows that one of the "ancient asexual scandals", Bdelloid rotifers (aka "water bears - they are microscopic but multicellular animals), show evidence of horizontal genetic transfer of genes. As asexual taxa are supposed to be unable to evolve as quickly as their predators and pathogens, this resolves a problem, along with the discovery that they have mutation resistant genes, as to how they could survive 80 million years or so without sex to mix and match useful genes. It turns out they use other taxa's genes. That fits nicely with my microbial species paper argument.
Finally, a paper that is being widely reported (e.g., by erv), in which the researchers tried to "run the evolutionary experiment" for a group of E. coli (Carl Zimmer's babies) more than once, to find, unsurprisingly, that evolution depended on contingent events (like rare mutations). This is being hailed a bit to show that Gould was right. Maybe he was, but I think there's a couple of philosophical points being glossed over.
One is that if a certain view of time is correct, the so-called "block universe" view, then only one evolutionary path was possible - the actual one. The contingency here arises from our lack of knowledge of the determinate causes, including on that view quantum events. The second is that every event in the experiment is still caused, one way or another, so the contingency here is that of a "chaotic deterministic system", a set of processes in which tiny variations magnify determinately over time to cause very different outcomes. Chaos theory, it is often not understood, is determinist (unless you set up statistical processes). So it may be that evolution simply had to do what it did on earth in those conditions; and this may not take away from the claim that evolution is sensitive to contingent events.
I'm not sure why I posted that but I think it's important to disambiguate the implications.
- Log in to post comments
interesting comparison about how catholics handle science and education versus how protestants do. I guess it's part of the back to fundamentals' policy they've had since Luther and Calvin.
I don't think I've heard a disagreement with Gould that was based on block time before, and I'm not sure how successful such an argument it is. Part of that has to do with epistemic problems regarding how far we can push our knowledge of an idea of time so far removed from human experience. I think that it's probably better to drop the block time component of the argument altogether and emphasize the deterministic aspect of chaos theory, and so illustrate the deterministic aspect of evolution.
Several of us have already discussed the paper since it was published online ahead of print back in May.
'Water bears' are tardigrades, not rotifers.
A minor point: rotifers are not "water bears", which is the English common name for tardigrades. The common name for rotifer is either "rotifer" or "wheel animalcule".
I knew I should have checked that first...
John,
OT,however....
PZ has just banned a certain troll from his blog,and there is much rejoicing(not that it bothers the troll,still posting),can I just say,even if you authorize all your regulars and all,thats not exactly the idea of a blog is it? I mean,I dont have the perfect solution,but Ive said this on pharyngula as well, if every sane and benevolent poster just ignored the trolls,they would have nothing to go on,and we wouldnt have to go through all this shit...Im really annoyed with this.
In the What is a Species? post on Panda's Thumb, I have left the trolls to their business, and it seem that you don't have to feed them for them to take over a thread (go see what I mean). For now, this is the best way to deal with it given the limitations of the technology...
The "block" version of time (eternalism, McTaggart's B-series) is claimed to be compatible with indeterminism and multiverse Everett-type scenarios, so multiple evolutionary paths should be possible. Apparently in this view, there are an infinite (or huge) number of static time blocks going out in all directions. It's just that there's no explanation for why "you" are in the block you're in. So the question would not be why evolution had to be a certain way, but why "you" ended up in this block. On the other hand, if you take the presentist view of time, then you have to explain "now", which seems unavoidably linked with consciousness. But then again, I can't see how you explain "now" in an eternalist view either.
Hmmm, I was interested in that paper until I read this:
"what the authors ultimately did was a cladistic analysis on 17 of the most complete hominin fossil skulls".
So that's n=17, total? Including all the contrasts and comparisions? Will somebody please teach the anthrologists some decent biology? Geometric morphometrics is very nice, but it doesn't generate statistical power out of thin air.
Re: Gould and contingency: It's funny you should mention them in the same breath as chaos theory. There's people who think a lot of Gould's more famous work can be explained by chaos theory. Power-curve distributions of extinction events vs. punctuated equilibrium, that kind of stuff.
Just a note on continuous characters: their use in cladistics is highly contentious. Continuous characters are not, by any stretch of the imagination, homologous. They are similarities, which is phenetic. I would call the analysis by González-José et. al. (2008, Nature 453, 775-778) phenetic, not cladistic.
The authors argue: "Theoretically, most cladistic methods need these characters to be discrete and independent, among other requirements". The characters need to be homologous! This is fundamental to cladistics! How can 0-33% be homologous in any sense?
I see that the paper is written by anthropologists, a group that has argued for the inclusion of continuous characters into cladistic analysis (see Rae 1998, Cladistics 14, 221-228). It is unfortunate that such a mediocre paper has made it into Nature. Sorry for being so narky, but this really irritates me!
I'm rather partial to the notion of "block universe" or "block time" or "timescape". Just aim your starship at the point in the sky where the Earth was 100 years ago, fire up the FTL warp drive and "presto!" you find the Earth of 1908 right where we left it. It means that all those people and - more importantly - pets that have passed away are really still out there somewhere, only dead from certain temporal perspectives. Getting back to them does pose certain technical problems but, think about it, it's almost as good as life after death.
I think it was Fred Hoyle's October 1st Is Too Late that got me interested in the idea and then Gregory Benford's Timescape that developed it by actually attempting an answer to the 'grandfather paradox' which most popular SF simply ignores.
This is in response to the comment post by MCE. Characters are not homologous... biological structures do. In this case, expressing the variation in a discrete or a continuous way is the same!!!
In fact, some (SEVERAL!) discrete characters, are just a continuous characters expressed in a different form.
And if you are so brilliant to say: "t is unfortunate that such a mediocre paper has made it into Nature", PLEASE, send a reply to Nature. I am sure it will be published.
Please, read Goloboff et al 2007 (Cladistics).