Friday, December 16, 2005

Endless forms most continuous   posted by Razib @ 12/16/2005 01:22:00 PM

What is evolution? I just had someone email me that people often ask him about his "belief" in evolution. This reiterates the point that people consider evolution a belief system, not a scientific paradigm. No matter how ridiculous it seems, how is it that we arrived at this juncture? There are multiple factors that have resulted in ~1/2 of Americans rejecting any form of evolution.

One important point is the reality that human beings as essentialists. That is, we imbue animals with essences. Elephants are elephants, even if you dress it up like a giant flamingo, there is something essential about elephants that is preserved. Paul Bloom would root this in our innate dualism. Others would elaborate in more detail aspects of innate derived folk biology. Children from Creationist and non-Creationist backgrounds tend to prefer narratives which imply immutable and essential natures to creatures all around them. As a point of plain fact that is what we see with our naked eyes. Whether or not evolution has shaped us toward a particular bias in regards to systemetizing about the natural world, the inputs we receive from the environment early on in development reinforce our conception of an ordered and discrete world where the essence of elephant is poured into the flesh of the proto-elephant.

This sort of thinking crops up in folk systematics (in mythologies about the beasts and birds and fish of the sea) as well as Creationist tracts which emphasize the importance of "kinds." Obviously Creationists have problems with hybrids like mules, and in particular, fertile hybrids like female tigons. But it is not only Creationists who have problems with female tigons, biologists who espouse the biological species concept (BSC) also have issues with the reality of fertile hybrides (this is why the peculiar, at least intuitionally, phylogenetic species concept emerged [PSC]).

The BSC was promoted by the famous evolutionary biologist Ernst Mayr. Mayr also was a thinker in an older school of evolutionary phylogenetics which relied on the intuition and knowledge of specialists in particular clusters of species. The problem is that the BSC, that a species is defined by the bounds of fertile matings across a population generating fertile offspring is not workable across vast swaths of the tree of life. Not only do asexual creatures fall outside of the testable criterion of viable crosses between two individuals (unless you count various forms of horizontal gene transfer), but many traditional species generate fertile hybrids, and amongst many plants hybrid viability is so omnipresent as to make a joke of the BSC. The problems with the BSC and evolutionary phylogenetics resulted in the rise of the cladists in the 1960s, who emphasize the use of synapomorphies in determining monophyletic clades. A mouthful, huh? Cladists and the simultaneous rise of phylogenetic trees were welcome revolutions in turning systematics toward a more deductive and falsifiable set of methods. But the problem is that the hypothetico-deducative methods that lay at the root of these new fields are not necessarily intuitive. That makes sense, as I have said multiple times, human operational thought is not deductive, but abductive, we work from a set of facts to a good working explanation, we do not work from a hypothesis to predictions which we rigorously test.

There are rough & ready concordances between the intuitive phylogenies that were produced by Mayr's school, and even Haeckel's Tree of Life. For that matter, conventional human creation stories generate decent evolutionary relationships as well, they are clearly not random fabulations, but derive from a solid grounding in the data on hand. The problem is that science is a precise and reproducible system, and gestalt knowledge can usually only be resolved by appeals to authority and loud shouts of "You're wrong, because I said so!" Despite the excessive fanaticism of cladists on shared derived characters, the rigor that they brought to systematics was certainly a good thing. And the challenge that they put forth to the BSC was also a good thing.

Because species are problems. Aside from species it is a traditional point of understanding that other taxonomical categories are arbitrary, i.e., genus (note that humans are not in the same genus as chimpanzees, though this does not make taxonomical sense). But even species can be problematic. This is an issue because the intuitions that we as humans have about species are, in my opinion, a major grounding for Creationist rejection of evolutionary theory. Once we dethrone species, intuition will no longer be the block it currently is. No longer will we have to appeal to the Talk Origins observed speciations page. No longer will we have to respond to the charge that there have been no observations of radical changes in body plan in one generation. No charges that "bacteria don't evolve." Ultimately these battles can be won, but the greater war is the sanctity of species, "kinds."

Plant geneticists like Loren Rieseberg of the University of Indiana have been studying the importance of hybridization amongst various "species" for many years. Not only does a retrofit of Biblical systematics stumble upon the ubiquity asexual microorganisms, it is a poor representation of evolutionary genetic dynamics amongst the primary producers of biogenic matter on this planet, the plants. Unlike animals various plant species hybridize promiscuously, and one could say that for some taxa the term species is simply shorthand for correlation structure across space and time, interlaced by a continuity of gene flow.

But how relevant is this for animals? Clearly animals do not hybridize as easily as oak trees or sunflowers. Nevertheless, essentialist truths collapse with the insertion of slim needles. Species concepts are build upon a house of cards which assumes that populations which are so demarcated are (magically) insulated from affinal populations by mating barriers, whether it be prezygotic (behavior, mating season, etc.) or postzygotic (lower fitness of hybrids, hybrid sterility, etc.). But the intuitional concepts are grounded (more or less) in an absolutist model, not a graduated one. Like fucks like.

In How species evolve collectively: implications of gene flow and selection for the spread of advantageous alleles Rieseberg and Moran explore the importance of interpopulational gene flow in maintaining a rough species cohesion. Their contention is that advantagenous alleles of large effect can spread through a worldwide population via selection even when migration is not sufficient to homogenize neutral alleles. In short, Reiseberg and Moran save the species concept by suborning the impression that a species' genetic architecture is a unitary and hypercontingent black-box which can not brook change or introgression. As I have noted before, the effect of gene flow on the distribution of frequencies across and within populations can be modeled with the equation 1/(4Nm + 1), where N is the population and m the proportion of migrants per generation. In short, Rieseberg and Moran argue that the rate of migration between populations within species is simply too low for genetic drift and local adaptation not to result in speciation. On the other hand, they argue that genes of large fitness effect may be able to spread even when the migration rates are not sufficient to equilibrate most of the genome. Recall that the probability of a mutation of fixation is 2s, and this can be translated into models which use migration in lieu of mutation to introduce genetic novelty. When s, the selection coefficient that defines the fitness of individuals with the allele vs. population mean fitness, is low, the probability of fixation is low, and random genetic drift is very powerful. But as s gets larger the probability of fixation increases and random genetic drift has a proportionately weaker role in the dynamics of frequency change. Rieseberg and Moran suggest that even on quantitative traits, where there are many loci, the locus of largest effect often explains a large fraction of variation, in which case the the alleles at these loci that confer great fitness will be able to sweep over worldwide populations and maintain species cohesion.

But the lessons to be learned from this paper extend beyond maintenance of putative species, they suggest the pluralism of microevolutionary mechanisms which are at work on the populational level. But we can take this further, in his paper The scale independence of evolution Armand Leroi expresses skepticism at the tendency for some to assert radically different dynamics on the macroevolutionary vs. microevolutionary level (i.e., clade level selection, etc.). He argues that there is a continuity between the two levels, microevolutionary forces of selection and drift are sufficient to explain taxonomical variety and novelty. This is an important point because it speaks to the idea that species and genus' are not privileged, and despite their operational intuitional appeal we may need to move beyond innate systematics to understand how evolution actually works as a process. Rieseberg above argued that spread of alleles of large positive selective effect tie together species, but other work by plant geneticists shows that such alleles can spread across "species." The essentialist concept of species is a fundamental block in the road to being open to this possibility. If species have peculiar essences sealed off from others it makes little sense that traits of one species would spread to another.

We privilege individuals and species as operational units because they are highly salient. Our minds are geared toward recognizing moving objects as individuals with agency. Faces are not simply smudges against the background, they stand out with particular detail. But in the broad scope of evolution individuals, and species, might not be as special as we believe. Certainly species are not intelligible in the context of asexual creatures beyond a gestalt level, and it is a highly problematic concept in plants as well. In concert with our tendency to think in terms of expanding non-reticulated genealogies, the species concept might be blocking our ability to understand the evolution of our own species. On the comment boards of this weblog someone expressed the idea that modern humans would simply not mate with neandertals because of physical differences. But the reality is that we know that humans have sex with ungulates. I believe what is happening is that we are abducing to the best possible explanation of the reality that we know we are special, and different. In their book Speciation H. Allen Orr and Jerry Coyne scoff at the possibility of human & non-human primate hybridization. This shows that even biologists are not immune from this bias when it comes to our own species. Orr's work in particular has shown that mutations which generate fitness benefits of large effect can spread rather fast, and operate outside the conventional infinte alleles models (which assume fitness differences of very small effect). Rieseberg uses Orr's work above as theoretical underpinnings for the paradigm that he is pushing forward.

In The Selfish Gene Richard Dawkins promoted a "gene centered" worldview that deemphasized individuals as "vehicles." This was in the context of microevolutionary theories promoted by W.D. Hamilton which derived the evolution of altruism from the realities of kinship and gene identity. But just as genes brook no differences of individuals, that is, we are simply vehicles that they use to traverse space and perpetuate themselves through time, so species may also be simply more difficult valleys to cross. Though sexual reproduction in our species is banal, when viewed across the span of eons, one can not ignore the possibility that interspecies matings could also be a process which is highly relevant to the evolution as selectively favored alleles that introgress might spread like wildfire. Genes are discrete units of information on the nucleotide level, but from their vantage point the existence of creatures of flesh and cellulose might simply be continuous scaffolds which present non-trivial, but not insurmountable, barriers to traversion. So, in answer to questions relating to speciation, I'm not sure how ground shaking these points really are in the grand scope of evolutionary theory.