Monday, May 14, 2018

A new yardstick



If we look at ancient DNA from 4,560 to 1,210 years ago, we see a steady increase over time in the number of genetic variants that are linked to high educational attainment (Woodley et al. 2017)



Four years ago I discussed genetic variants that seem to favor high educational attainment (Frost 2014). They’re found at single nucleotide polymorphisms (SNPs), and their incidence varies from one human population to another. In all but one case, they are specific to humans and not shared with ancestral primates. 

Davide Piffer has been interested in these SNP variants, seeing them as a possible way to measure how genes contribute to intelligence in different populations. By looking up population data, he can calculate their average incidence for a given group of people. This measure is called the “cognitive polygenic score.”

When he wrote up his latest paper (Piffer 2017a), only nine of these variants were known. For each geographic region, the scores were as follows:

Sub-Saharan Africans – 18%
Amerindians – 25%
North Africans – 30%
Oceanians (Papuans, Melanesians) – 34%
Southeast Asians – 35%
West Asians – 38%
Middle Easterners – 40%
Europeans – 41%
Siberians – 43%
East Asians – 45%

This regional breakdown is open to criticism. Sardinians (32%) were not included in the European category, and Mongolians (49%) were grouped with East Asians rather than with Siberians. The distinction between Middle Easterners and West Asians is not clear to me. The Amerindian category is based on a few small groups. And who is included in the Southeast Asian category? Only Cambodians?

When Piffer compared these scores with the results of IQ tests in these regions, he found a high correlation of 0.9. That is high, higher than what I would expect, given the quality of the data, especially for mean IQ, and the very disparate nature of the two datasets.

Over a two-year period Piffer submitted his paper to Intelligence, resubmitted it, had it rejected, and then resubmitted it to Frontiers in Psychology, where it was accepted by the reviewers before being rejected by the editor. It is now sitting in the limbo of a preprint repository (Piffer 2017a).

Meanwhile, the number of these SNPs has continued to grow. A research team led by Aysu Okbay identified 74 SNPs that are associated with educational attainment (Okbay et al. 2016). Another team led by David Hill reported 107 in their initial preprint and 187 in their published paper (Hill et al. 2018). 

Piffer (2017b) repeated his analysis, now using the 107 SNPs that Hill’s team had identified. The geographic pattern still held up but was weaker, the correlation being only 0.64. This lower score is actually more in line with what I would expect. It diverges the most from mean IQ in two geographic areas:

1. South Asia (Pakistan, India) - South Asians seem to do worse on IQ tests than their genetic endowment predicts. Why? Is it the culture? The diet? Inbreeding? Perhaps language. IQ tests are often administered in a language (English, Hindi, Urdu) that may be the second language of the person taking it. Or perhaps South Asian educational attainment is determined not only by IQ but also by qualities like the ability to sit still and not make a ruckus in class.

2. The Mende of Sierra Leone - For some reason, the Mende have a higher cognitive polygenic score than any other African population. This might be a real finding, or a typo.

Another research team, led by Michael Woodley, has compared the Okbay dataset with ancient DNA to see whether the cognitive polygenic score has increased over time, specifically between 4,560 and 1,210 years ago. The DNA was retrieved from European sites and a few sites from southwest and central Asia. The result? The cognitive polygenic score did increase over time. People on average had more and more of the alleles that favor educational attainment. The authors note that IQ alone may not be responsible:

[...] While the increase in these variants over time is certainly consistent with the expectation of rising GCA [general cognitive ability], the possibility that their increase indicates a simultaneous rise in other factors that make unique contributions to educational attainment (such as 'slow' life history or 'high-K' social cognitive characteristics) cannot be ruled out. (Woodley et al. 2017; references within quote removed)

The new mental/behavioral package developed through a process of feedback with the cultural environment. This gene-culture coevolution likely continued into recent times:

This process likely continued until the Late Modern Era, where it has been noted that among Western populations living between the 15th and early 19th centuries, those with higher social status (which shares genetic variance with, and is therefore a proxy for GCA) typically produced the most surviving offspring. These in turn tended toward downward social mobility due to intense competition, replacing the reproductively unsuccessful low-status stratum and effectively 'bootstrapping' those populations via the application of high levels of skill to solving problems associated with production and industry, eventually leading to the Industrial Revolution in Europe. (Woodley et al. 2017; references within quote removed)


Conclusion

More and more SNPs are being linked to educational attainment. The total is now in the triple digits. That’s still less than the thousands of genes that influence intelligence, but there is no need to identify most of them to spot general trends. Selection acts on phenotype, not on genotype. Selection for intelligence should therefore impact all of these SNPs in the same direction. It’s like estimating the proportions of different colors in a bowl of Smarties. You don’t have to count every last one. Just pick out a handful at random and count the colors.

Four years ago only 7 SNPs had been linked to educational attainment. Now we have 187. In another four years we’ll probably have more than a thousand. All the same, I doubt that the overall geographic pattern will change much. The problems lie elsewhere:

-          Genetic data may be lacking for some unmixed groups, particularly Amerindians.

-          The relationship between intelligence and cognitive polygenic score may not be linear.

-          We may be relying too much on educational attainment as a proxy for IQ (which itself is a proxy for intelligence).

When I was in public school, girls did better than boys in almost every subject. They had good attendance, always took notes, and did their homework. Boys got bored more easily and spent more time fidgeting, daydreaming, and drawing pictures in their notebooks. This sex difference exists in all cultures, but it seems greater in some than in others.

How useful is educational attainment as a proxy for IQ? Yes, these two measures correlate highly with each other (Rindermann 2018, pp. 51-54), but this high correlation is based on studies from WEIRD countries (Western, educated, industrialized, rich, and democratic). Does it hold up on a global scale? I’m not so sure.


References

Frost, P. (2014). Population differences in intellectual capacity: a new polygenic analysis, Evo and Proud, March 8
http://evoandproud.blogspot.ca/2014/03/population-differences-in-intellectual.html

Hill, W. D., R.E. Marioni, O. Maghzian, S.J. Ritchie, S.P. Hagenaars, A.M. McIntosh, C.R. Gale, G. Davies, I.J. Deary. (2018). A combined analysis of genetically correlated traits identifies 187 loci and a role for neurogenesis and myelination in intelligence. Molecular Psychiatry
https://doi.org/10.1038/s41380-017-0001-5

Okbay, A., J.P. Beauchamp, M.A. Fontana, J.J. Lee, T.H. Pers, C.A. Rietveld, et al. (2016). Genome-wide association study identifies 74 loci associated with educational attainment. Nature 533: 539-542.
http://www.nature.com/articles/nature17671

Piffer, D. (2017a) Evidence for Recent Polygenic Selection on Educational Attainment and Intelligence Inferred from GWAS Hits: A Replication of Previous Findings Using Recent Data. Preprints, June 8
https://www.preprints.org/manuscript/201706.0039/v1

Piffer, D. (2017b). Piffer's results replicated (again) by latest GWAS (N=147,194), toppseudoscience, July 21
https://topseudoscience.wordpress.com/2017/07/21/piffers-results-replicated-again-by-latest-gwas-n147194/comment-page-1/#comment-95

Rindermann, H. (2018). Cognitive Capitalism. Human Capital and the Wellbeing of Nations. Cambridge University Press.

Woodley, M.A., S. Younuskunju, B. Balan, and D. Piffer. (2017). Holocene selection for variants associated with general cognitive ability: comparing ancient and modern genomes. Twin Research and Human Genetics 20(4): 271-280.
https://doi.org/10.1017/thg.2017.37 

19 comments:

Sean said...

I am afraid a fair proxy for IQ would be professed disbelief in IQ differences between groups. According to Robin Dunbar, Plains Indians had a tradition of low born warriors becoming “war chiefs” who would never retreat in battle and usually got quickly killed although they had a slight chance of rising to the top. Race differences is a very risky area, and the gatekeepers of journals probably assume such research is a gambit by bad scientists with little to lose.

China's long standing Imperial exam system must have resulted in their real world IQ being less impressive than their hereditary adaptation to education achievement might suggest. The quota to limit their representation at elite universities may not be more justified that is commonly thought. Another people may be underrated, and merit their supposed Ivy League over representation.

Peter Frost said...

"a fair proxy ... would be professed disbelief" - I read this sentence three times, and the logic still escapes me.

"Another people may be underrated, and merit their supposed Ivy League over representation" - Same problem.

sykes.1 said...

The trend line in Woodley's figure (shown above) is determined by the two most recent data. The earlier data show no trend.

This looks like a typical social science garbage plot with no substantial significance despite a low p-value.

That said, the differences in socio-economic outcomes between races, especially when they are seen in different geographic regions, are so obvious that substantial difference in IQ's between races cannot be doubted.

Anonymous said...

What African groups were analyzed?

Sean said...

A good proxy for IQ would be the ability to get on in the world, which is what IQ is there for. The most intelligent people of all will tend to express views about racial/genetic differences in intellectual attainment that their superior intellectual ability has determined will be a means to the end of getting on in life. In other worlds, the cleverest people will see through the pretense that they are supposed to tell the truth, and say what society will reward them for saying. So equaliterianism and racial egalitarianism are a proxy for very high IQ. The most intelligent people of all will tend to take it a step further and exert themselves to come up with plausible theories that explain why IQ environmentalism is true, like Ron Unz does.

"China's long standing Imperial exam system must have resulted in their real world IQ being less impressive than their hereditary adaptation to education achievement might suggest. The quota to limit their representation at elite universities may be more justified that is commonly thought. Another people may be underrated, and merit their supposed Ivy League over representation".

I was trying to say that Chinese are probably less clever that their high number of genes for passing exams, and their actual exam results, might suggest, and so limiting their representation at from top universities is justified.

"Another people" meant Jews, I think they are underrated by IQ measures and actually merit their supposed Ivy League over representation. The Chinese were selected by Eight Legged Essay writing, while traditional Jewish higher education involved arguing a case in the complex argot of Talmudic disputation. Jews are to Chinese as boys are to girls.

Anonymous said...

I don't think there was much selection by the imperial exam system. The number of people who took and passed the exams relative to the rest of the population was very small. And the typical candidate spent years in schooling, taking exams and acquiring degrees, delaying marriage and family formation, similar to grad students today. There may have been negative selection for successful exam takers. Also the Japanese are similar and they never had an exam system.

Regarding real world success, I don't think there's a great mystery there. In addition to a higher mean IQ, Jews have different personality and behavioral traits. And Jews have co-evolved for a long time as a caste among Europeans dealing with European personality and behavioral traits. As an urban caste, their "natural environment" consisted of Europeans and their behavioral and mental traits. Whereas the Chinese have been small scale peasant farmers. Obviously the Jews' evolutionary history is more relevant to contemporary success in modern European and European derived post-industrial societies.

Peter Frost said...

"The trend line in Woodley's figure (shown above) is determined by the two most recent data."

Sykes,

If only two of the 66 data points were driving that trend, it wouldn't be significant. In fact, the correlation between the polygenic score and year is significant, p=.04

Anon,

Biaka, Mbuti, Yoruba, Madenka.

Sean,

"A good proxy for IQ would be the ability to get on in the world"

"Getting on in the world" depends on many things, and not simply intelligence. It usually depends on being physically attractive, having the right social connections, and holding the right political views. And all of that can vary from one place to another. During the Cambodian genocide, intellectuals were targeted for death. To a lesser extent, this was true during the Cultural Revolution in China and the Stalinist terror of the 1930s. Marshal Zhukov would talk like a peasant to ward off suspicions of being too smart for his own good. In the U.S., President Eisenhower used an "Aw shucks" persona to win people's trust.

From my personal observation, very smart people often don't get on the world. Above an IQ of 140, intelligence is often associated with social awkwardness and emotional immaturity. A 140+ IQ is not shaped by natural selection; it's a product of developmental accidents that may have negative side-effects. It's the same thing with an IQ of less than 80. Really smart and really stupid people act strange.

Up to a certain level, intelligence is associated with the ability to assimilate mainstream ideology. Beyond that level, one starts to ask too many questions and entertain too many doubts. If you really want to get on in life, it's best to be a bit smart but not too smart.

"Chinese are probably less clever that their high number of genes for passing exams, and their actual exam results, might suggest, and so limiting their representation at from top universities is justified"

I've been told by Chinese people that they're good at assimilating information but not so good at examining information critically --- "thinking outside the box."

"I think they are ..."

Oh s---. Here you go again!! Your opinion is less important than your ability to defend it. Please give references to support your view.

Anonymous said...

Here's an interesting graph:

http://www.unz.com/isteve/asians-are-most-realistic-about-white-black-iq-gap/

Asians perceive the white-black IQ gap to be significantly higher than most other groups do, while Jews perceive it to be significantly lower than most other groups do.

If we define "intelligence" as the ability to perceive objective reality and to hold and express beliefs in accordance with such perceptions, then Asians would be, at least on this issue, more "intelligent".

But of course there are many other possible interpretations. It could just mean that they are more honest on this issue. Or it could mean that they are actually less intelligent by holding the less socially acceptable opinion. Or it could mean that they are just more socially inept. It could mean that they are good at observing and assimilating facts from their life experience, but lower in the critical thinking ability that allows one to override such observations with complex cogitations. Or it could mean that they have the critical thinking ability to override social propaganda which promotes egalitarian views.

Similar such interpretations and questions could be raised regarding the Jewish response.

Anonymous said...

Something to look forward to -
GWAS meta-analysis (N=279,930) identifies new genes and
functional links to general cognitive ability

Human Genomics 2018, 12(Suppl 1):A88
Background
Cognitive ability is highly heritable [1] and a major determinant
of human health and well-being [2]. Recent genome-wide metaanalyses
have identified 24 genomic loci linked to variation in
general cognitive ability [3-7], but much about its genetic
underpinnings remains to be discovered. Here, we present the
largest genetic association study to date (N=279,930) of cognitive
ability.
Materials and methods
We performed a genome-wide meta-analysis of 16 independent cohorts
of European ancestry. Cognitive ability was assessed using
various measures of performance in fluid domains of cognitive functioning.
Despite sample and methodological variations, genetic
correlations (rg) between cohorts were considerable (mean=0.63),
and there was no evidence of heterogeneity between cohorts in the
single nucleotide polymorphism (SNP) associations. GWAS was
conducted following a standardized quality control procedure, including
e.g. checks on missingness, genotyping errors, relatedness
and corrections for possible population stratification. Meta-analysis
was carried out in Metal.
Results
We detected 12,701 genome-wide significant variants, represented
by 246 independent lead SNPs in 206 genomic loci (191 novel).
Roughly half of the significant SNPs are located within a gene, and
4.4% are highly likely to have a regulatory function. The results implicated
1041 genes via positional, eQTL and chromatin interaction
mapping, and gene-based analyses. 89 genome-wide significant hits
were exonic nonsynonymous variants, with clear implications on
gene function. Stratified LD Score regression showed enrichment for
heritability of SNPs located in conserved regions of the genome, coding
regions, H3K9ac histone regions, and super-enhancers. While
conserved regions have been implicated for cognitive ability and
neuropsychiatric phenotypes, enrichment in coding regions has not
previously been linked to variation in cognitive ability. Pathway analyses
showed independent associations of three biological pathways
(neurogenesis, neuron differentiation and synaptic structure), several
brain tissues, and specific brain cell types (cortical and hippocampal
pyramidal neurons, and striatal medium spiny neuron cells). We
confirm previously reported genetic correlations with several neuropsychiatric
traits, and used Mendelian randomization to test for
credible causal associations. We found a strong protective effect of
cognitive ability on ADHD and Alzheimer’s dementia, bidirectional
causation and strong pleiotropy between cognitive ability and
schizophrenia, and a risk effect of cognitive ability on autism.
Conclusions
This study provides the most extensive meta-analysis of genetic variants
associated with cognitive ability to date, provides novel insight
into the underlying neurobiology and into causal relationships with
other traits. The results are a rich resource of functionally relevant outcomes,
which can provide starting points for further functional work.

Sean said...

John Derbyshire (in The Marx of the Anti Semites) writes."Anyone who has read stories from the premodern period of China’s history knows that the guy who gets the girl—who ends up, in fact, with a bevy of “secondary wives” who are thereby denied to less intellectual males—is the one who has aced the Imperial examinations and been rewarded with a District Magistrate position. This went on for two thousand years."

Wikipedia on the exam "The eight-legged essay was formulated around a rigid, artificial structure. It tested, among other things, the examinees' knowledge of the Four Books and Five Classics and ability to insert classical allusions and idioms at the places deemed appropriate. The structure of much of the essay included heavy parallelism and redundancy, rhetorical features that survive in modern Chinese expository writing."

According to On Their Own Terms, by Benjamin A. ELMAN "By 1600 Europe was ahead of Asia in producing basic machines, such as clocks, levers, and pulleys, that would be necessary for the mechanization of agriculture and industry."

In IQ tests the Chinese have different strengths to Europeans, who are stronger on verbal tasks. In addition to being somewhat higher in IQ, Jews are almost a SD stronger than Gentile Europeans on IQ test verbal tasks. The good stuff seems to go with verbal strength, and certainly not with the peculiarly Chinese intellectual profile. And that is why I think Chinese students are overrated, except in the narrow area of being successful students.

But even if Chinese are actually rather underrated, and as all rould clever on average as their 9% extra "cognitive polygenic score" genes suggests, their enrollment must be restricted at the top universities. Unless one wants it widely known that the top universities of the West have an absurdly low proportion of Gentile Europeans, Jews or Chinese numbers must subject to a glass ceiling. For given that a Chinese facility for hard and unremitting attention to their studies leads to them getting far higher examination marks than others with the same capacity for the innovatory acumen (what top universities must foster) the key point is that everyone can see who is Chinese, while over-representation of Jews is not remotely as obvious. Without the policy that Ron Unz asserts is actually already entrenched at elite universities, an ethnic minority would visibly dominate dominate top universities' enrollment by virtue of meritocratic over-representation. And that would be a fundamentally unstable state of affairs.

Anonymous said...

The terminology of studying the differences between populations needs to change, especially in light of the advancements in genomic research, which places the racial labels of "white", "black" and "asian" quite out of court. Population genomics demonstrates 5 highly clustered, geographical populations in the human species. These populations correlate to some extent with racial typologies, but not suitably for scientific purposes given the variability in social perception of racial types, as mentioned by David Reich.
For instance, Europeans do not constitute their own population, rather they are members of West Eurasians ( loosely correlates with the category 'Caucasoids') a group which includes Middle Easterners, North Africans and some Central/South Asians. West Eurasians share an ancestral connection to farmers and hunter-gatherers in the Middle East.

The point being, how productive is it to compare groups like Chinese and Euros to the exclusion of their entire genetic cluster? What can you infer about the cognitive propensities of East Asians from looking at Chinese, Koreans and Japanese alone? Surely this inflates the representative IQ as low IQ nations such as Laos, Cambodia, Thailand, Phillipines are not included in such representations?

I submit that IQ is not the best proxy for the solely genetic propensities of a population as there is bound to be large fluctuation in scores over 1000's of years, given dysgenic variables such as war and eugenic variables such as agricultural surplus. If IQ demonstrates intelligence, and the latter results in innovative civilisations, surely a country like Iraq would demonstrate an IQ on par with Europeans seeing as Iraq (Ancient Mesopotamia) is the birthplace of the earliest civilization in human history, not to mention the written language, wheel, mathematics, concept of time among others? Or Egyptians for that matter?
Population wide analyses would be more prudent in such studies.

Anonymous said...

The exam system isn't 2,000 years old. It's generally dated from about 650 AD, and until the Ming Dynasty (1368–1644), the numbers seem to have been very small. And the numbers don't seem to have been that great during the Ming and Qing dynasty (1644-1904) years:

https://en.wikipedia.org/wiki/Chinese_examination_system

"The success rate on these tests of knowledge on the classics was between 10 and 20 percent, but for the thousand or more candidates going for a jinshi degree each year in which it was offered, the success rate for the examinees was only between 1 and 2 percent: a total of 6504 jinshi were created during course of the Tang dynasty (an average of only about 23 jinshi awarded per year).[25]"

"Statistics indicate that the Song imperial government degree-awards eventually more than doubled the highest annual averages of those awarded during the Tang dynasty, with 200 or more per year on average being common, and at times reaching a per annum figure of almost 240.[32]"

"The examination system was revived in 1315, with significant changes, during the reign of Emperor Renzong. The new examination system was one of regionalism with Mongol characteristics. The northern areas of Mongolia and its vicinity were favored, and a quota system (both for number of candidates and number of degrees awarded) which was based on the classification of the imperial population into four racially-based groups (or castes and/or ethnicities) was instituted, the groups being Mongols, their non-Han allies (Semu-ren), Northern Chinese, and Southern Chinese, with further restrictions by province.[35] Under the revived and revised system the yearly averages for examination degrees awarded was about 21.[36]"

"Near the end of the Ming dynasty, in 1600, there were roughly half a million licentiates in a population of 150 million, that is, one per 300 people; by the mid-19th century the ratio had shrunk to one civil licentiate for each 1,000 people."

https://www.princeton.edu/~elman/documents/Civil%20Service%20Examinations.pdf

"The civil service examination system, a method of recruiting civil officials based on merit rather than family or political connections, played an especially central role in
Chinese social and intellectual life from 650 to 1905."

"Civil service examinations by themselves were not an avenue for considerable social mobility, that is, they were not an opportunity for the vast majority of peasants and artisans to move from the lower classes into elite circles. The archives recording data from the years 1500 to 1900 indicate that peasants, traders, and artisans, who made up 90 percent of the population, were not a significant part of the 2 to 3 million candidates who usually took the local biennial licensing tests . "

Anonymous said...

I don't know who underrates Jews, except perhaps a small minority of anti-Semites. Most anti-Semites don't underrate Jews.

I don't think the imperial exam system and Talmudic disputation explain the difference between Chinese and Jews. The exam system only involved a small minority of the Chinese population. Moreover, the Japanese are similar and never had such a system. And as Cochran has shown, the Jews evolved as an urban caste of white collar workers in Europe.

The Jews have a higher IQ, different personality and behavioral traits, and co-evolved with Europeans as an urban caste. While the Chinese evolved as peasant farmers. I think these explain the differences in real world success. The Japanese have a similar evolutionary history as the Chinese do. Different personality traits determine success in Japan compared to the US:

https://www.sciencedirect.com/science/article/pii/S0889158318300054

" Agreeableness significantly contributes to men's annual income in Japan.


The income premium effect is strengthened at large firms in Japan.


Agreeableness acts as a wage penalty for male workers in the U.S."

Peter Frost said...

"Something to look forward to - GWAS meta-analysis (N=279,930) identifies new genes and
functional links to general cognitive ability"

We're going to see a lot of new data over the next two to three years. Yet the overall patterns will not change much.

"If we define "intelligence" as the ability to perceive objective reality and to hold and express beliefs in accordance with such perceptions, then Asians would be, at least on this issue, more "intelligent"."

Intelligence is the ability to process thoughts. There is no guarantee that those thoughts are valid. Historically, many intelligent people -- and in fact entire civilizations -- have fallen prey to mass delusions of one sort or another. On this specific point, there are historical reasons why Jewish people support antiracism, even to the point of becoming delusional and refusing to hear counter-arguments. That isn't lack of intelligence.


"How productive is it to compare groups like Chinese and Euros to the exclusion of their entire genetic cluster?"

The number of human populations is arbitrarily large or small. It depends on the level of resolution you wish to impose on the data. For some purposes, it is useful to treat Chinese, Koreans, and Japanese as a single population. They share a common heritage of Buddhism and, to a lesser extent, Confucianism. They have advanced state formation. They have influenced each other, even in the words they use, and so on. For other purposes, it is useful to look at them separately. For other purposes still, it is useful to look at subgroups, like the Burakumin.

"surely a country like Iraq would demonstrate an IQ on par with Europeans seeing as Iraq (Ancient Mesopotamia) is the birthplace of the earliest civilization in human history"

Much has changed in Mesopotamia over the past six thousand years. It's not just that the population has changed (influx of Arabs and African slaves). The population has also undergone different regimes of natural and social selection.

"The numbers [of Mandarins] seem to have been very small."

The Mandarins were those who passed all three levels of the civil service exam. There were social benefits, however, even for people who passed the first level (i.e., at the local prefecture). Such passage provided exemption from labor service and corporal punishment, government stipends, and admission to upper-gentry status.

The base of the exam pyramid broadens even further if we include everyone who prepared for the first level. If the 3rd level graduates were a fraction of the 1st level ones, the latter were likewise a fraction of all Chinese children who studied with a view to taking the exam. Test-taking, selection, and elimination occurred even at this early family stage.

Anonymous said...

The base of the exam pyramid broadens even further if we include everyone who prepared for the first level. If the 3rd level graduates were a fraction of the 1st level ones, the latter were likewise a fraction of all Chinese children who studied with a view to taking the exam. Test-taking, selection, and elimination occurred even at this early family stage.

The paper I linked to above discusses local level tests:

"Civil service examinations by themselves were not an avenue for considerable social mobility, that is, they were not an opportunity for the vast majority of peasants and artisans to move from the lower classes into elite circles. The archives recording data from the years 1500 to 1900 indicate that peasants, traders, and artisans, who made up 90 percent of the population, were not a significant part of the 2 to 3 million candidates who usually took the local biennial licensing tests."

At the bottom of the exam period were the "licentiates", those who passed the first level of tests at the local level and eligible to take exams at the county level:

"Government control of civil and military selection quotas was most keenly felt at the initial licensing stages for the privilege to enter the examination selection process at the county level. In 1600 China had perhaps 500,000 civil licentiates in a population of 150
million, or a ratio of 1 licentiate per 300 persons. By 1850, with a population of 350 million, China had only 800,000 civil and military licentiates, but still only about a half-million were civil, a ratio of 1 per 1,000 persons."

According to the paper, 90% of the population generally were not part of the 2 to 3 million who took the first level of local tests. Of this 2 to 3 million, 500,000 to 400,000 passed and became "civil licentiates", which gave them the privilege to take county level tests.

So about 0.3% to 0.1% of the population were "civil licentiates" at the base of the exam pyramid. For comparison, the percentage of Americans with a PhD is about 2%.

Peter Frost said...

We may be talking past each other. What matters is the proportion of the population that participated in the exams, and that proportion is 10%. Families allocated resources to those children who showed promise, and that sort of resource allocation influenced other life decisions, including marriage. All of this is prior to the actual exam.

A proportion of 10% doesn't seem like much, but it can influence overall genetic change in a population, especially in a population where most people fail to reproduce. Ron Unz discusses this point at length in his article: "How social Darwinism made China":

Each generation, a few who were lucky or able might rise, but a vast multitude always fell, and those families near the bottom simply disappeared from the world. Traditional rural China was a society faced with the reality of an enormous and inexorable downward mobility: for centuries, nearly all Chinese ended their lives much poorer than had their parents.

Because of the marked shortage of women, there was always a great number of men without wives at all. This included the overwhelming majority of long-term hired laborers… The poorest families died out, being unable to arrange marriages for their sons. The future generations of poor were the descendants of bankrupted middle and rich peasants and landlords.

In a Guangdong village, a merchant family named Huang arrived and bought land, growing in numbers and land ownership over the centuries until their descendants replaced most of the other families, which became poor and ultimately disappeared, while the Huangs eventually constituted 74 percent of the total local population, including a complete mix of the rich, middle, and poor.

http://www.unz.com/article/how-social-darwinism-made-modern-china-248/

Anonymous said...

Around 10% were involved in the first stage of exams, and about 0.3% to 0.1% passed the first stage. Presumably just passing the first stage conferred some benefits, but the numbers seem too small to have made much of an impact. A lot more, more than 9%, failed to pass. Are you suggesting that simply participating, taking the test and failing, conferred benefits and was selected for? Perhaps it did. I don't know enough about Chinese history to have an opinion on that. But in that case, the tests and successful test taking ability wouldn't be doing the selecting. Having the means and wherewithal to be able to take the test or have your children take the test would be selected for. The tests would be a kind of luxury good and test taking would be a form of conspicuous consumption.

I'm familiar with Unz's paper. He makes the same point, that the genetic impact of test taking would have been negligible:

"With Chinese civilization having spent most of the past 1,500 years allocating its positions of national power and influence by examination, there has sometimes been speculation that test-taking ability has become embedded in the Chinese people at the biological as well as cultural level. Yet although there might be an element of truth to this, it hardly seems likely to be significant. During the eras in question, China’s total population numbered far into the tens of millions, growing in unsteady fashion from perhaps 60 million before AD 900 to well over 400 million by 1850. But the number of Chinese passing the highest imperial exam and attaining the exalted rank of chin-shih during most of the past six centuries was often less than 100 per year, down from a high of over 200 under the Sung dynasty (960-1279), and even if we include the lesser rank of chu-jen, the national total of such degree-holders was probably just in the low tens of thousands,[12] a tiny fraction of 1 percent of the overall population—totally dwarfed by the numbers of Chinese making their living as artisans or merchants, let alone the overwhelming mass of the rural peasantry. The cultural impact of rule by a test-selected elite was enormous, but the direct genetic impact would have been negligible."

Peter Frost said...

Again, we're talking past each other. Yes, the number of people who passed the exams ("degree holders") was small, but the number of people participating in the exam culture was large: 10%. Perhaps 10% seems like a small number, but it's large in terms of selection pressure, particularly in a country where most men historically failed to reproduce.

Anonymous said...

I don't think 10% is a small number, but if the overwhelming majority of that 10% failed and merely participated in the exam culture, then successful test taking presumably wasn't selected for. If just participating in the tests and being part of that 10% were selected for, then it seems like whatever allowed one to be in a position to take the tests, like being wealthy and having leisure, would have been selected for instead.