Let George Do It

I was thinking about how people would have adapted to local differences in essential micronutrients, stuff like iodine, selenium, manganese, molybdenum, zinc, etc. Australia, for example,  hasn’t had much geological activity in ages and generally has mineral-poor soils. At first I thought that Aboriginals, who have lived in such places for a long time,  might have developed better transporters, etc – ways of eking out scarce trace elements.

Maybe they have, but on second thought, they may not have needed to.  Sure, the Aboriginals were exposed to these conditions for tens of thousands of years, but not nearly as long as kangaroos and wombats have been.  If those animals had effective ways of accumulating the necessary micronutrients,  hunter-gatherers could have solved their problems by consuming local fauna. Let George do it, and then eat George.

The real problems should occur in people who rely heavily on plant foods (European farmers) and in their livestock, which are generally not adapted to the mineral-poor environments. If I’m right, even in areas where sheep without selenium supplements get white muscle disease (nutritional muscular dystrophy), indigenous wildlife should not.

 

Posted in Australian Aboriginals, Dietary adaptations | Tagged | 70 Comments

More on Deafness

Last time, I was talking about the  hearing children of deaf parents.  They could give us some interesting information on the effect of a verbally impoverished environment in early life , which is the currently fashionable explanation for low IQs and low academic achievement in some non-Parsee minority groups.  But since there have been very few decent studies of these kids, we don’t have much info.  At least I haven’t found much.

On the other hand, we have quite a bit of info about kids who are themselves deaf, and of course they have an even more impoverished verbal environment.

Deafness has a big impact.  Average verbal IQ is 85, a standard deviation below normal. Nonverbal IQ is  normal – in the case of deaf children raised by deaf parents, who have the most exposure to sign language, it may be higher than  normal.  Note that these deaf-of-deaf  kids are almost entirely of European ancestry (96%), probably because of the common 35delG connexin-26  mutation, which likely gives heterozygotes some advantage.

Blacks in the US have a similar average verbal IQ, but also score lower on nonverbal IQ tests.  In fact, their disadvantage is greater in nonverbal IQ than in verbal IQ.

Seems to me that limited verbal stimulation is not a very plausible primary cause of low test scores and low academic achievement in blacks, because  the degree of deprivation needed to cause a 1-standard deviation decline is extreme (deafness), and because there is an even greater depression of nonverbal scores, which, judging from the results in deaf children, should not be affected at all by limited verbal stimulation.

 

 

 

 

Posted in Uncategorized | 116 Comments

Neanderthal Man: In Search of Lost Genomes

Svante Pääbo has a book out, Neanderthal Man, in which he recounts his adventures sequencing ancient DNA.  He has had three big successes: the first successful sequencing of Neanderthal mtDNA in 1997, the first sequencing of the Neanderthal nuclear genome, and later the first Denisovan genome.

Ancient DNA is usually very degraded: short DNA sequences mixed with bacterial DNA, and often contaminated by modern human DNA.  Pääbo and his team made major contributions in sample preparation and sequencing methodology. The interpretation of that data has been performed by people like David Reich, Nick Patterson, and Monty Slatkin – and a good thing too, because Pääbo is no theorist. At each stage of his work, he had certain expectations about the results, and those expectations were nearly always wrong.  He thought any Neanderthal genetic contribution to modern humans was very unlikely – but it’s there.  To be fair, that was the case for many other people working in human genetics. I’ve never really understood why.

Or later, after a few-percent Neanderthal admixture had been shown to exist in people outside of sub-Saharan Africa, he at first thought that it probably had no functional consequences.  He believed that the correct null hypothesis was that a genetic change would have no consequences whatever: but that’s silly.  The question wasn’t whether one particular Neanderthal allele was advantageous, but whether any of them were.  In order for his null model to be correct, Neanderthals would had to be inferior indeed, not better adapted to their home territories in any way.  Although again, to be fair, there are whole branches of science in which the favorite null model is always wrong.

One interesting side point: by looking at the entrails of the online supplement to the big Neanderthal paper in May 2010, it was possible to see that there was something odd about Melanesians: they were genetically more distant from Africans than other Eurasians.  Which implied another dose of archaic ancestry. Judging from this book (which may not have the complete story)  the people working that problem didn’t notice that anomaly, but instead compared against the just-sequenced Denisovan genome and noticed that Melanesians were significantly closer to Denisovans than  other Eurasians.

A good experimentalist can (sometimes)  get to some reasonable approximation of the truth even if he’s short on theory.  Worth remembering, particularly in the human sciences, where emotions make theory gang aft a-gley.

 

Posted in Archaic humans, Book Reviews, Denisovans, Genetics, Neanderthals | 114 Comments

Heteropaternal Superfecundation

Sperm competition is a factor in some cases of paternal uncertainty, but there are also many cases in which it is not. For example, Daniel Boone’s wife had a daughter, Jemima, when he had been gone for more than a year.  There is reason to believe that Jemima was fathered by Daniel’s brother Edward. Non-paternity, but no sperm competition.  Adoption also creates non-paternity without any sperm competition. Generally speaking, sperm competition ought to be more likely in a species that has a seasonal estrus cycle – all effective mating attempts have to be close together in time.

In a case where there is sperm competition, we can assume that the non-official parent fertilizes half the time. This means that 2 times the frequency of non-paternity is an upper limit on the fraction of conceptions in which sperm competition occurs.

Since every recent, high-quality set of genetic measurements shows that non-paternity is currently low and has been low for hundreds of years, usually under 2%, for those populations that have been studied,  the fraction of conceptions in which sperm competition occurs is less than 4% (in those places and times). We have an upper limit – but is it a good upper limit?  Is it close to the actual rate?

Once in a while, women give birth to fraternal twins that have different fathers. The frequency of this event can give us a direct measure of the fraction of conceptions in which sperm competition actually occurs.  In the presence of sperm competition, we only notice the cases in which the twins are discrepant.  The incidence of sperm competition will be twice that fraction.

So what is the rate?  This is the sort of question that could be answered definitively nowadays with a SNP chip, but here’s what we have: In an Italian study from 1992,  the frequency was  2.4% among fraternal twins whose parents were involved in paternity suits – people who presumably had reason to be suspicious. Generally speaking, nonpaternity is much more common in high-suspicion cases of this sort, usually at least 10 times more common than in the general population.  Another 1992 estimate was that 1 in 400 pairs of twins born to married white women had different fathers, which give a sperm competition rate of about half a percent – reasonably compatible with the Italian numbers.

Maybe we need to redo this with modern genetic techniques. Maybe Italian girls are special.  Maybe everything was different back before recorded history.

But if these numbers are correct and at all representative, sperm competition in humans is insignificant, and all the people talking about our specialized somatic and behavioral adaptations for sperm competition are wrong.

Someone mentioned that I am calling people loons who’ve authored of dozens of peer-reviewed papers and are at serious (sic) universities.

I can only say – somebody’s gotta do it!

 

Posted in Genetics | 119 Comments

Simple Mobility Models II

Sweden

This is a sequel to the previous post exploring quantitative genetic models of income inequality motivated by the findings in Gregory Clark’s new book [1].

A simple quantitative genetic model must rely on the assumption of an underlying normal distribution of something, EQ we can call it, that is the additive genetic part of whatever determines income. We can impute this since, now, we have no way to measure it. Like IQ originally, and like our construction of AQ,’amish quotient’, here, we could come up with an estimator if we could measure what we think we ought to measure. For the moment we assume that there is such a direction in character space. An immediate problem is that income is far from normally distributed but we can impute a mapping from the observed income distribution to EQ. Income percentiles are well known and published for many countries. The figure is derived from Swedish data given by Björklund and Jäntti [2].

The the top panel shows the conventional Lorenz curve for the Swedish data: the horizontal axis is income rank and the vertical axis is the percentile of national income at that rank, i.e. it is a conventional cumulative distribution of income. The circles show data points along a green line which is a spline through the data points.

The bottom panel has the same vertical axis but the distribution along the horizontal axis is the imputed normal distribution of EQ. For example the 50th percentile of income maps to the mean of the imputed distribution, the 84th percentile maps to +1 standard deviation of a standard normal, the 16th percentile maps to –1 standard deviation, and so on. This figure gives us estimate of the income of a person given his value of an underlying normally distributed EQ, shown as superimposed on the figure. The virtue of this is that we can instantly apply a century’s worth of quantitative genetic theory and knowledge.

In each panel a computed offspring distribution is shown as a bar two standard deviations wide with a line to the parents’ EQ at each mid-quintile, i.e. at percentiles 10,30,50,70,and 90. The lines are not vertical because of regression to the mean.

Given a quantitative genetic model we know, for example, that offspring of a couple should be distributed symmetrically around the mid-parent value, regressed toward the mean.
A clear exposition of this is in a post in Steve Hsu’s blog , with a contribution from James Lee, here. This particular figure is computed with an additive heritability of 0.9 and a spousal correlation of 0.9. These are both at the high end of plausibility, as suggested by Clark’s data. We also should remember that family (cultural) transmission, of wealth or values or whatever is likely indistinguishable from genetic transmission and will increase the heritability.

From this we can compute, and perhaps derive explicit expressions, for the long term movement, i.e. EQ, of one’s descendants. It should not be difficult to derive longer term expectations from a model like that in the bottom panel. For example starting at some initial EQ the distibution of descendants’ EQ should be distributed in subsequent generations along the EQ distribution, the X axis of the bottom panel, according to some reaction-diffusion process similar to the Fisher-Komogorov equation where the diffusion is given by the the incomplete assortative mating and the heritability less than unity. The deterministic part would describe the greater number of surviving offspring with high EQ as documented in Clark’s Farewell to Alms [3].

To paraphrase what we hear often in the social sciences, there is no need to invoke any social or cultural transmission at all.


  1. Clark, G., 2014. The Son Also Rises: Surnames and the History of Social Mobility. Princeton University Press.  ↩
  2. Intergenerational income mobility in Sweden compared to the United States. A. Björklund and M. Jäntti.
    The American Economic Review (1997):1009–1018.  ↩
  3. Clark, G., 2007. A Farewell to Alms. Princeton: Princeton University Press.  ↩
Posted in Uncategorized | Tagged , , | 37 Comments

Simple Mobility Models

Greg’s previous post reviewing Gregory Clark’s new book has generated some interesting discussion. Reviewers elsewhere and some of our customers have been surprised at the persistence of social class by surname and speculated that this implies genetic transmission of class. Does it?

A way to pursue the issue is to construct the simplest possible competing models and to compare their predictions. This is nearly the universal procedure in all of science. First we can construct a pure social science model in which class i simply culturally transmitted. A typical paper in the literature (these days about income) divides a sample into quintiles and describes (agonizes, often) about mobility between income quintiles. We can shorthand these, calling families under the twentieth percentile “lower”, the next quintile “lower middle”, then “middle”, then “upper middle”, then “upper.” We might find, then, that 15% of the sons of lower class families are lower middle class the next generation, and so on.

Cultural Transmission

At this point some pop social scientists and journalists rely for interpretation on a covert assumption (or sleight of mind) that this is equivalent to a transition between states in a Markov process, that is that particles (i.e. families) forget their pasts. Our 15% transition rate from lower to lower-middle is an estimate of the probability of jumping to lower middle given a middle class family in one generation. Now we may see that the corresponding transition probability from from lower-middle to middle is, say, 10%. If the social structure is static we immediately deduce that the probability that a lower class surname is middle class after two generations is simply the product, 0.15 * 0.10 or 1.5%. Under this kind of model one’s status becomes independent of that of the more distant ancestor and randomly distributed across classes, assuming that all classes are reachable, eventually, from all other classes.

The justification for the Markov assumption is, in our toy model, that class is purely culturally transmitted some someone from a lower-middle class family’s progress is determined by cultural transmission of lower-middle class culture plus some error term that leads to switching class. We do not assume family transmission, just class transmission, because family transmission would just mimic genetic transmission. [This needs worked out, since there is no explicit model of what cultural transmission really is that we know about.]

Genetic Transmission

A simple quantitative genetic model must rely on the assumption of an underlying normal distribution of something, EQ we can call it, that is the additive genetic part of whatever determines income. We can impute this since, now, we have no way to measure it. Like IQ originally, and like our construction of AQ (‘amish quotient’) a few weeks ago, we could come up with an estimator of it if we could measure what we think we ought to measure. For the moment we assume that there is such a direction in character space. An immediate problem is that income is far from normally distributed but we can impute a mapping from the observed income distribution to EQ. Income percentiles are well known and published for many countries. The figure below shows data given by Björklund and Jäntti (1997).

Sweden

The top panel of the figure shows the conventional Lorenz for the Swedish data: the horizontal axis is income rank and the vertical axis is the percentile of national income at that rank, i.e. it is a conventional cumulative distribution of income. The bottom panel has the same vertical axis but the distribution along the horizontal axis is the imputed normal distribution of EQ. For example the 50th percentile of income maps to the mean of the imputed distribution, the 84th percentile maps to +1 standard deviation of a standard normal, the 16th percentile maps to -1 standard deviation, and so on. This figure gives us estimate of the income of a person given his value of an underlying normally distributed EQ. The virtue of this is that we can instantly apply a century’s worth of quantitative genetic theory and knowledge.

Given a quantitative genetic model we know, for example, that offspring of a couple should be distributed symmetrically around the mid-parent value, if heritability is complete, with standard deviation of sqrt{2}/2. If not, we add regression to the mean. If mating is assortative we add that to the model. We can go one and on adding bells, whistles, and coontails to the model but we start simple.

From this we can compute, and perhaps derive explicit expressions, for the long term movement, i.e. EQ, of one’s descendants. In other words we can test this model against the kind of data that Gregory Clark has gathered and falsify, or not, the genetic model.

We can see right away that the pure cultural model (our version) is falsified since status persists. Our genetic model of diffusion along an EQ axis shows that the Markov assumption is far from satisfied. If someone is at EQ 0.91, corresponding to the 82nd percentile, toward the lower end of the upper class, his offspring are much more likely to fall to the upper-middle class than is someone at EQ 1.5, corresponding to the 93rd income percentile. On the other hand, those fallen offspring are much closer to the upper class threshold that are offspring who entered the upper-middle class from the middle class. They are much more likely to diffuse back up a generation later. Once we lump people into quintiles that Markov property vanishes along with the pop-social-science interpretations of mobility statistics.

Since this is the sort of model that is familiar to physicists, I think we can agree that Greg owes it to us to work out more of the details.

References

Intergenerational income mobility in Sweden compared to the United States. A. Björklund and M. Jäntti. The American Economic Review (1997):1009–1018.

The Son Also Rises: Surnames and the History of Social Mobility. G. Clark. Princeton University Press (2014).

Posted in Uncategorized | Tagged , , , , | 34 Comments

The Son Also Rises

Greg Clark has a new book out, The Son Also Rises. His thesis, in short, is that moxie has high heritability. Most studies show fairly high social mobility from one generation to the next – but Clark finds (using surname analysis) that paternal lineages that were over-represented (or under-represented) in measures of status (such as education or wealth) can be still be over-represented (or under-represented)  in high-status groups several hundred years later.  Both statements are true.  In the short run, from one generation to the next, luck plays a big role.  In the longer run, the fact that the subpopulation being examined has a different genotypic average, one more likely to result in high status, means that regression to the mean of the general population is slow for the subgroup, essentially caused by gradual change in its average genotype, change produced by intermarriage with individuals who on average have a less favorable genotype.  Other than high heritability, the other prerequisite for this pattern is highly assortative mating for moxie. If two groups have different average amounts of moxie, complete endogamy (as in Indian castes) would ensure that the between-group difference would continue indefinitely,  disregarding selection.

We All Are Tall!

Here’s a simple example.  Take a group of NBA players – they’re a good deal taller than average.  Assume that they all marry WNBA players ( it’s a thought experiment, ok?). The kids will be taller than average, and probably some get into the NBA (far more than average) , but most don’t. There’s a lot of change in status from one generation to the next.  Have them continue to marry among themselves: they stay that tall, and each generation is  over-represented in the NBA. In any generation, kids in a particular family   are gaining or losing NBA status, but people in this clan are regressing to a higher mean.

Now instead, imagine that 10% marry out each generation.  The people they marry are probably not as as tall, but even more important,  they are, on average,  genetically shorter. Keep up this admixture for generations and our NBA/WNBA clan will eventually converge to the population mean.

Where did  Clark and his students see this this pattern of slow long-term social mobility?

Everywhere they looked. England, Sweden, Japan, Korea, China, Chile.  In England, Norman surnames are still 25% over-represented at Oxford and Cambridge, but then it’s only been 947 years. The Japanese upper class is something like half Samurai (5% of the population when they lost their special privileges, 143 years ag0).

Assuming that is in fact a genetic phenomenon, such slow convergence also requires pretty low paternal uncertainty – but then paternal uncertainty is in fact low, at least in the populations we have looked at. I talked to Clark about this: he hadn’t really looked at it yet and had heard that the rate was around 10%.  Where do people get these notions?  Now he knows better.

It’s not in the book (ongoing work) but it turns out that the long-term pattern is the same if you look at matrilineal descent. As it should be, if it’s genetic.

It turns out that you can predict a kid’s social status better if you take into account the grandparents as well as the parents – and the nieces/nephews, cousins, etc. Which means that you’re estimating the breeding value for moxie – which means that Clark needs to read Falconer right now. I’d guess that taking into account grandparents that the kids never even met, ones that died before their birth, will improve prediction.  Let the sociologists chew on that.

Adoption – turns out most of the kid’s status is due to genetic factors, rather than family environment.  At least for reasonably normal environments: this may not be the case if you’re raised in a barrel and fed through the bung-hole.

Often groups with a different average genotypic value are generated by a process of biased leaving and/or joining – like upper and lower classes or the Amish.  Natural selection can also do this, in an endogamous group.   Of course, you can do the same thing by importing a population that already has a higher or lower average.  As long as it doesn’t mix much, it can stay different (higher or lower) for a long time.

If culture was the driver, a group could just adopt a different culture (it happens) and decide to be the new upper class by doing all that shit Amy Chua pushes, or possibly by playing cricket. I don’t believe that this ever actually occurs.  Although with genetic engineering on the horizon,  it may be possible.  Of course that would be cheating.

It is hard to change these patterns very much. Universal public education, fluoridation, democracy, haven’t made much difference.  I do think that shooting enough people would. Or a massive application of droit de seigneur, or its opposite.

Clark finds that windfalls don’t make much difference in the long run. Back in 1830, they kicked the Cherokee out of Georgia and distributed the land by lottery in 1832. One-fifth of the adult male white Georgians were winners, with a value of something like $150,000 in 2014 dollars.  But by 1880, their descendants were no more literate, their occupational status no higher.  Sounds like modern lottery winners, or NBA players, yes?  The major exception must be extreme poverty: a windfall that keeps you from starving to death must have long-range effects on your descendants.

If moxie is genetic, most economists must be wrong about human capital formation.   Having fewer kids and spending more money on their education has only a modest effect: this must be the case, given slow long-run social mobility. It seems that social status is transmitted within families largely independently of the resources available to parents. Which is why Ashkenazi Jews could show up at Ellis Island flat broke, with no English, and have so many kids in the Ivy League by the 1920s that they imposed quotas.  I’ve never understood why economists ever believed in this.

Moxie is not the same thing as IQ, although IQ must be a component. It is also worth remembering that this trait helps you acquire  status – it is probably not quite the same thing as being saintly, honest, or incredibly competent at doing your damn job.

Posted in Ashkenazi Jews, assortative mating, Book Reviews, Education, Genetics | 72 Comments