Wednesday, March 21, 2012

Evolution Not Very Settled

In The Selfish Gene, Richard Dawkins argued that it is not species that compete, not even individuals, but genes. You'd never learn it from Dawkins, but it appears selection happens at all sorts levels, not just at the gene level. In Haidt's Righteous Mind, he (a psychologist) gives the following example of group selection.

A geneticist worked with cages containing twelve hens each, and he picked the cages that produced the most eggs. Then he bred all of the hens in those cages to produce the next generation. Within just three generations, aggression levels plummeted, and within 6 generations death rates fell from 67% to 8%, and eggs produced jumped from 91 to 237. If you just picked the hens that laid the most eggs, they tended to be the meanest, and so aggressive behavior (and dead hens) went up. That is group selection working better than individual selection, and gene selection is even less relevant; there is selection of genes, but only at the level of the group.

Then there's Steven Jay Gould's famous statement that 'there's been no biological change in humans in 40k to 50k years.' Haidt notes no one believes that anymore, especially after this Russian guy, Dmitri Belyaev, turned wild foxes into harmless dogs in only 30 generations, or about 600 years in human terms. That means the invention of agriculture 8k years ago could have really changed our brains and behavior at least as significantly as how foxes differ from dogs. It follows pretty basically that human races are probably truly different at some levels (mainly similar, to be sure).

These aren't minor changes. You have to pick this stuff up on the periphery, in blogs like gnxp, or books by Cochrane and Harpending, not the mainstream evolution books, which tend to focus on refuting intelligent design and creationists.


Anonymous said...

Not that your point isn't correct, but Belayev killed all but the least aggressive pups, leaving very few, very select individuals in each generation. This kind of non-random selection is very strong and will fix characteristics much more quickly than selection in the wild. It's not clear how strongly selection has acted on humans. Some has been strong in local populations (e.g. genocides), but fixing traits across the whole population in 8k years? I don't think we've yet seen evidence for Belayev-strength selection in humans since the invention of agriculture. But there is evidence of a "bottle neck" much earlier.

Eric Falkenstein said...

On the other hand, he merely selected the pup that came closest to him. So, it was a pretty noisy signal

B. A. said...

pinker has an interesting section on the possible contribution of recent genetic changes to the general decline in violence.

Anonymous said...

Even Dawkins conceded that group selection of the type observed in the hen experiment was theoretically possible. In the Selfish Gene he talked about the conditions required to be satisfied - groups of less 25 individuals and less than one individual transferring between groups each generation (that's from memory - I'm reasonably confident I've got the right book). These conditions are, of course, satisfied in the hen experiment. It is the absence of evidence for group selection outside of experiments that is the cause of its general dismissal.

As for where the debate is now (if you don't take the blogging world as evidence of that debate), it is generally accepted that is possible to examine systems at multiple levels ("multi-level selection" is the term of the moment). However, selection at one level can be mathematically transformed to be at another level, so ultimately, can be explained in terms of gene level selection (i.e. inclusive fitness). The practical question is which level of analysis yields the most insight? There is a great video where Stuart West poses that question and suggests that the contribution of group selection to modern evolutionary theory is roughly zero:

As for the human question, you are spot on with the potential for recent human evolution. And on that note (for a blatant plug), the economic effects of recent human evolutionary change is the subject of my research and much of my blog: (although that page does need an update)

Mercury said...

"These aren't minor changes. You have to pick this stuff up on the periphery, in blogs like gnxp, or books by Cochrane and Harpending, not the mainstream evolution books, which tend to focus on refuting intelligent design and creationists."
Thats because its also political nitro glycerin. But at some level, asseting that human evolution stopped dead in its tracks ~50k yrs. ago (as our species began to separate into geographically isolated groups)isn't any more plausible than intelligent design/creationsim.

Dipper said...

My crude understading is that early man was faced with a major problem in that water carries harmful bacteria. Europeans used alcohol based drinks to counter this, and East Asians boiled water and used infusions.

One view is that different responses to alcohol (Europeans generally more alcohol tolerant) are a result of evolution. My uninformed view was that there hadn't been time for these differences to be selected in different populations, but maybe not.

Razib Khan said...

I don't think we've yet seen evidence for Belayev-strength selection in humans since the invention of agriculture

lactose tolerance in many northern european populations is nearly swept to fixation.

i think haidt is a little too convinced of higher-than-individual-level selection. but humans are a better case i think than other organisms.

Eric Falkenstein said...

If you read this Razib, I'd be really interested if someday you respond to Bowles and Gintis's post on the fitness of IQ. Given his math, if he's right, I don't see how any trait can flourish. That is, if genetic IQ explains so little in income inequality given its correlation with income, how did we develop language? I doubt our incidental increases in language ability were much greater in fitness than our IQ benefits, and this happened in a mere 100k-ish years. It seems mathematically, if Bowles and Gintis are correct, it takes a bizarre bottle-neck to create modern humans (hopeful monsters, I guess).

Eric Falkenstein said...

BTW, I read your earlier response on B&G, but didn't see how this jibed with evolution in general.

Mat Stroller said...

Maybe you will be interested in Cochran & Harpending's blog: West Hunter.

Razib Khan said...

That is, if genetic IQ explains so little in income inequality given its correlation with income, how did we develop language?

iq is considered a quant trait. language *competency* (not fluency) is not. so they're different in nature first off. second, i doubt the distribution of fitness of a trait at time x are always the same as at time x - t or x + t. the fact that iq is normally distributed is a strong hint that it hasn't been subject to strong *uni*-directional selection in a while. finally, if iq is polygenic you might be seeing fitness trade offs at the upper tail due to the g-matrix and correlated response.

Jason Collins said...

Having posted my comment about Dawkins's views above, I went back to The Selfish Gene and my memory on his use of the particular group size was incorrect - I must be thinking of his writing somewhere else. The more general point he makes:

"Genetically speaking, individuals and groups are like clouds in the sky or dust-storms in the desert. They are temporary aggregations or federations. They are not stable through evolutionary time. Populations may last a long while, but they are constantly blending with other populations and so losing their identity. They are also subject to evolutionary change from within. A population is not a discrete enough entity to be a unit of natural selection, not stable and unitary enough to be 'selected' in preference to another population."

I should note that Adaptation and Natural Selection by George Williams, a major influence behind The Selfish Gene, goes through the theoretical requirements and their likelihood in some depth (particularly in Chapter 4).

On IQ, as Razib suggests, the heritability of a trait and its phenotypic effects in modern populations is not evidence of past effects, as selection tends to eliminate variation. There is also strong evidence of selection of genes associated with brain size over short periods:

Then there are the assumptions that Bowles and Gintis go which drive much of the result:

Anonymous said...

Evolution is the ability to have a lot of kids and have them live long enough to have more. You guys ascribe a religious aspect to it which fits into your social Darwinist philosophy of wealth=fitness. Look around the world at the populations who are growing --this doesn't seem to be too accurate.

In fact evolution is a rather mundane process which doesn't do much to explain -- in the sense it can predict -- how species end up the way they do. It is an ex post facto science, which in reality gives it very little value beyond its ability to debunk the fairy tales of religions.
Science is measured by how well laws can predict the future, this is how Einstein was recently tested regarding neutrinos.