April 15, 2011

Origin of Language in Southwest Africa (?)

Recently, I've presented some conflicting evidence about the origin of modern humans in either East, or Central, or South Africa. A new paper looks at the decay of phonemic diversity in world languages, coming to the conclusion that language itself may have originated in Southwest Africa.

I will only point out two caveats:
  1. We cannot assume that click speakers of the African Southwest are necessarily indigenous to that region, and
  2. It is possible that, the greater phonemic diversity is due to ancient admixture between quite divergent peoples who possessed two different types of phonemic inventories, while most Africans inherited only the phonemic inventory of one of these peoples, which then decayed as per the author's theory away from Africa.
In fact, I suspect that something much like #2 (the Afrasian-Palaeoafrican hypothesis) probably took place. I'll leave an evaluation of this paper to the experts, but it seems to me that South Africans do not only have a richer phonemic inventory, but also a set of phonemes (clicks) that seems peculiar to them, and may, possibly be part of a different language system than the one that most other languages use.

Science 15 April 2011:
Vol. 332 no. 6027 pp. 346-349
DOI: 10.1126/science.1199295

Phonemic Diversity Supports a Serial Founder Effect Model of Language Expansion from Africa

Quentin D. Atkinson

Human genetic and phenotypic diversity declines with distance from Africa, as predicted by a serial founder effect in which successive population bottlenecks during range expansion progressively reduce diversity, underpinning support for an African origin of modern humans. Recent work suggests that a similar founder effect may operate on human culture and language. Here I show that the number of phonemes used in a global sample of 504 languages is also clinal and fits a serial founder–effect model of expansion from an inferred origin in Africa. This result, which is not explained by more recent demographic history, local language diversity, or statistical non-independence within language families, points to parallel mechanisms shaping genetic and linguistic diversity and supports an African origin of modern human languages.

Link

38 comments:

German Dziebel said...

I posted my critical comments on this paper at Razib's: http://blogs.discovermagazine.com/gnxp/2011/04/the-african-ur-language/

But if we take the study at face value, what I find intriguing, in the light of the Out of America hypothesis, is that, according to this study, the best-fit second origin region is South America (Suppl. Mat Fig. S6), which is precisely the area of world highest language diversity measured in terms of independent stocks. Amazonia is the area with the greatest number of language isolates, which are also small populations, hence have a tendency to maintain smaller phonemic inventories.

batman said...

Which of theese languages have the most developed sound-system - in terms of range, complexity and distinction of sounds available to the human tongue?

If we allow homo erectus the rigth to speak - what is then left of this study?

ashraf said...

I think the linguistic diversity in south america has to do with isolation while in Africa, probably, the greater diversity was obscured by expansion of Nilo-Saharan, Afrasan and Niger-Kongo languages

German Dziebel said...

"I think the linguistic diversity in south america has to do with isolation while in Africa, probably, the greater diversity was obscured by expansion of Nilo-Saharan, Afrasan and Niger-Kongo languages"

This makes sense. But what the expansions of Khoisan, Nilo-Saharan, Afroasiatic and Niger-Congo languages did in Africa is they drove the intragroup genetic diversity of African populations up through effective population size growth and continent-wide intermixture. Hadza is the remnant of the original Paleoafrican landscape: limited genetic diversity, high divergence, just like New World or Melanesian populations. So, at the end of the day Africa used to look like the New World. The prediction of the out of African model is the opposite: the whole world used to be like Africa (genetically diverse but linguistically homogeneous) but subsequently it lost this diversity through serial founder effects. Out of America explains archaeology, genetics and linguistics much better.

batman said...

Sometimes I Like To Think,

I think. Thus the question of "Phonemic Diversity" can be a matter of identifying the premises for the term itself.

We like to think that nature is rational, in principe. Talking about "phonemic identities" and "phonemic diversities" we need first to define historical cicumstances were some of our ancestors experienced A conscious NEED to develop the phonetical range and qualities of their oral expressions.

Why someone start to create a wider range of morphemes and phonemes must be explained first - than we may find a logic to which we can compare the various levels of expression that occur in the various languages we know. But we can't start by guess-working how a "proto-altaic" could have looked like - on paper - if we don't have a clue to their basic sounds and their respective semantic.

Otherwise we need a cathegory for 'Assumed languages' that future students should learn to refrain from...

Beforee we have a better concept of the basic qualties of the human sounds of communication we still have no classifying concept that may tell us what the "the most simple sounds" may be.

Btw.: Since our 'common sense' are different from the 'common sense' of ancient antiquity, it gives little or no sense to adress the issue of "primitive sounds" without a prior and due termonology.

Explaining click-sounds to be "primitive" or "foraging" is - so far - not more than a good example of how easy it is to create presumptions out of mere asumtions.

Without other objective facts about the development of language, there is no way we can determine the original cause, motivation and purpose of the click-sounds.

How Does It Sound?

Our spoken languages can be identified as "sound-systems". Today we can analyze sound in all its aspects. On that basis we have been understanding the various historical backgrounds of each language-family. Thus we already know that an evolution of these languages have taken place over a very long time.

Today it is possible to determine the major language-families, that consists of continental "family-lines", regional "family-clusters" and local families. Though - when we compare the large family-lines with each other there is no - repeat NO - linguistic science that can define one to be older than the other. Exept by assumption...

To define structures, formats and encodings in one and the same matrix - to produce a possible time-line - we need a true historic relation, which only 'family-lines' - where the relationship is proven by semantics.

There is no way you can prove which language is the oldest - from Malay, Chineese, Indian, Polynesian, Aborignin and Zuni or Hopi. In fact we can't even say which of the family-line from Altaic, Uralic or German that is the older. So far we have no linguistic method that discriminate the various qualities of sounds available to the human ear and toungue - as a basis for a research into the phonetic realms of the fenotypes that appear out of the FOX2 genotypes.

Finally a word on "diversity":

To merly count limbs and branches of ONE three - and than compare this slavicly to other family-three - is just about bogus. What biologist will count the number of branches, limbs and leaves to find out if the oaks are older than the elms - as a specie...?!

Before the linguists show up with a sound-check that can match their spell-checking they will continue to analyze "preliminary presumtions" - as in 'premisses given by other professions'.

ashraf said...

there are some non khoisan people who use click sounds to call dogs or donkeys and also some people use such sounds to mean "no" without forgetting babies that produce such similar sounds

Onur said...

The prediction of the out of African model is the opposite: the whole world used to be like Africa (genetically diverse but linguistically homogeneous)

The Out of Africa theory has no prediction about linguistics, as that would extremely be a speculative and futile task. The fact that Africa is still the genetically most diverse place on earth despite the genetically very significant agriculturalist and pastoralist expansions there speaks volumes about how even more genetically diverse Africa was before those expansions.

Eugene said...

Amazonia is the area with the greatest number of language isolates
Really? I read so many times that New Guinea has the most unrelated languages, has that changed?

Kepler said...

I also think New Guinea may be the most varied linguistic landscape nowadays. I am not sure how it was earlier and I wonder if there is a really objective way of measuring language diversity.

In Southern Venezuela, just South of the Orinoco, we still have several Carib languages, several Yanomamö (completely different language family), two from the Saliba-Piaroa language family, some Pume, Guahibana, Macro-maku. Two centuries ago there were many more. In Brazil there are 20 different language families.
The differences between such Carib languages as Pemon and Yekuana or Panare are very important and seem to show a split well before this era. Panare even became a tonal language.

On the other side, it's almost impossible to determine the speed of language evolution in such an environment as the Amazon jungle, where rivers were by far the main paths to move around.

German Dziebel said...

@Eugene

"Really? I read so many times that New Guinea has the most unrelated languages, has that changed?"

I don't know where you read it. But the truth is that the New World harbors 2/3 of world language diversity. It contains 140 language stocks. Amazonia has the largest number of isolates (individual languages with no group affiliation whatsoever) and small families (2-3 languages max), which is the bottomline of linguistic diversity estimates. Papua New Guinea has more language families per unit of geography, but by the absolute number and by the number of isolates New World is unparalleled. Plus the New World likely suffered undocumented language extinctions since 1492.

A good online article is here www.rogerblench.info/.../New%20World/Peopling%20of%20the%20New%20World%20Canberra%20paper.pdf

Be that as it may, the fact that it's the New World and Papua New Guinea that show the highest levels of linguistic diversity, and not Africa, flies in the face of Atkinson's application of the out of Afrcia with serial bottleneck idea to the linguistic data.

@Onur

"The Out of Africa theory has no prediction about linguistics, as that would extremely be a speculative and futile task. The fact that Africa is still the genetically most diverse place on earth despite the genetically very significant agriculturalist and pastoralist expansions there speaks volumes about how even more genetically diverse Africa was before those expansions."

The idea behind the out of Africa theory is that populations were losing genetic diversity as they were colonizing the globe. So, Asians once were as diverse as Africans because they were Africans, Amerindians were once as diverse as Asians because they were Asians, etc. Linguistically, however, Africa and Europe are the two least diverse continents, which people explain as resulting from agricultural and pastoral expansions in the past 10,000 years. It means that Africa and Europe used to be as diverse linguistically as America and Papua New Guinea. But the genetic effect of agricultural and pastoral expansions would be an increase - not a decrease as you seem to imply - in genetic diversity through intermixture and demographic growth. It means that what we usually interpret as a sign of African antiquity is likely a product of the last 10,000 years. America preserves the original Mid-Late Pleistocene hominid condition (low demographic population size, low effective population size, low population density, high levels of population isolation and divergence measured by Fst, high levels of linguistic diversity because of isolation and lack of intermixture) better than Africa.

"Africa is still the genetically most diverse place on earth"

Not true. After 1492 the New World has grown to become the most genetically diverse place on earth. When geneticists are using African Americans as proxies for Africans, they are in fact sampling New World genetic diversity.

German Dziebel said...

@Kepler

"I also think New Guinea may be the most varied linguistic landscape nowadays. I am not sure how it was earlier and I wonder if there is a really objective way of measuring language diversity."

See above abut Papua New Guinea. Linguistic classifications have been handcrafted over the past 200 years. Genetic phylogenies are much more recent, machine-based and sweeping. Genetic code entails only 4 possible states (A, G, T, C), with a much greater likelihood of T and C and A and G interchanging, whereas there are many more possible states in units of linguistic analysis. The greater variability of linguistic code ensures that if the correct formal matches have been identified (no easy task, that's why scholars spend time weeding out borrowings and look-alikes) they are not homoplastic. So overall linguistics is, IMO, at least as objective as genetics. There are a number of well-studied cases when linguistics predicts population ancestry better than some genetic systems. Roma in Europe, Burushaski in South Asia, Ket in Siberia, Munda in India, etc. - all of them showed significant gene flow with neighbors that obscured their true genetic affiliations. Linguistically, their ancestry or the status of an isolate had been established long before mtDNA studies came of age.

See, e.g., Friedlaender, J., Hunley, K., Dunn, M., Terrill, A., Lindström, E., Reesink, G., & Friedlaender, F. (2009). Linguistics more robust than genetics [Letter to the editor]. Science, 324, 464-465.

German Dziebel said...

My previous comment didn't go through. It said the following:

@Eugene

"Really? I read so many times that New Guinea has the most unrelated languages, has that changed?"

I don't know where you read it. But the truth is that the New World harbors 2/3 of world language diversity. It contains 140 language stocks. Amazonia has the largest number of isolates (individual languages with no group affiliation whatsoever) and small families (2-3 languages max), which is the bottomline of linguistic diversity estimates. Papua New Guinea has more language families per unit of geography, but by the absolute number and by the number of isolates New World is unparalleled. Plus the New World likely suffered undocumented language extinctions since 1492.

A good online article is here www.rogerblench.info/.../New%20World/Peopling%20of%20the%20New%20World%20Canberra%20paper.pdf

Be that as it may, the fact that it's the New World and Papua New Guinea that show the highest levels of linguistic diversity, and not Africa, flies in the face of Atkinson's application of the out of Afrcia with serial bottleneck idea to the linguistic data.

@Onur

"The Out of Africa theory has no prediction about linguistics, as that would extremely be a speculative and futile task. The fact that Africa is still the genetically most diverse place on earth despite the genetically very significant agriculturalist and pastoralist expansions there speaks volumes about how even more genetically diverse Africa was before those expansions."

The idea behind the out of Africa theory is that populations were losing genetic diversity as they were colonizing the globe. So, Asians once were as diverse as Africans because they were Africans, Amerindians were once as diverse as Asians because they were Asians, etc. Linguistically, however, Africa and Europe are the two least diverse continents, which people explain as resulting from agricultural and pastoral expansions in the past 10,000 years. It means that Africa and Europe used to be as diverse linguistically as America and Papua New Guinea. But the genetic effect of agricultural and pastoral expansions would be an increase - not a decrease as you seem to imply - in genetic diversity through intermixture and demographic growth. It means that what we usually interpret as a sign of African antiquity is likely a product of the last 10,000 years. America preserves the original Mid-Late Pleistocene hominid condition (low demographic population size, low effective population size, low population density, high levels of population isolation and divergence measured by Fst, high levels of linguistic diversity because of isolation and lack of intermixture) better than Africa.

"Africa is still the genetically most diverse place on earth"

Not true. After 1492 the New World has grown to become the most genetically diverse place on earth. When geneticists are using African Americans as proxies for Africans, they are in fact sampling New World genetic diversity.

Onur said...

Linguistically, however, Africa and Europe are the two least diverse continents, which people explain as resulting from agricultural and pastoral expansions in the past 10,000 years.

That is a completely valid explanation in light of the known facts.

But the genetic effect of agricultural and pastoral expansions would be an increase - not a decrease as you seem to imply - in genetic diversity through intermixture and demographic growth.

No, agricultural and pastoral expansions brought homogenization and loss of genetic diversity (in addition to loss of linguistic diversity) to Africa.

Not true. After 1492 the New World has grown to become the most genetically diverse place on earth.

We aren't talking about the post-1492 New World, so that's irrelevant.

When geneticists are using African Americans as proxies for Africans, they are in fact sampling New World genetic diversity.

Geneticists haven't been using African Americans as proxies for Africans for a very long time and are instead have been using contemporary Sub-Saharan Africans, and the highest genetic diversity is in Sub-Saharan Africa argument still holds.

German Dziebel said...

"That is a completely valid explanation in light of the known facts."

Good we agree on that. But mind you the magnitude of language replacement in Sub-Saharan Africa means that the foraging populations must have been rather sparse and not as linguistically diverse as their counterparts in the Americas or Papua New Guinea.

"No, agricultural and pastoral expansions brought homogenization and loss of genetic diversity (in addition to loss of linguistic diversity) to Africa."

Eh? Genetic diversity is a function of population size. The invention of agriculture has brought about a cumulative increase in population size and the allelic diversity has gone up with it. Agriculture and pastoralism also resulted in population mixture on the continent-wide scale that also drove genetic diversity up. Haplotypic diversity and pairwise nucleotide differences are consistently higher in African agriculturalists than in African hunters and gatherers. Bantu "diverged" earlier than Pygmies. Take a look at Fig. 2 in "Phylogeography of the human mitochondrial L1c haplogroup: Genetic signatures of the prehistory of Central Africa" by Batini et al.

"We aren't talking about the post-1492 New World, so that's irrelevant."

No, that's very relevant. The post-1492 colonization of the New World is a normal population process and a good example of how genetic diversity has nothing to do with the age of a population. The New World is now more diverse than Sub-Saharan Africa because it sampled populations from every corner of the globe.

Onur said...

There are a number of well-studied cases when linguistics predicts population ancestry better than some genetic systems. Roma in Europe, Burushaski in South Asia, Ket in Siberia, Munda in India, etc. - all of them showed significant gene flow with neighbors that obscured their true genetic affiliations. Linguistically, their ancestry or the status of an isolate had been established long before mtDNA studies came of age.

That's not true, as linguistics and genetics give answers to very different questions and only genetics can give indisputable answers regarding ancestry and true genetic affiliations.

German Dziebel said...

"That's not true, as linguistics and genetics give answers to very different questions and only genetics can give indisputable answers regarding ancestry and true genetic affiliations."

On the individual level, yes, genetics can assign you to a unique cluster but on the group level linguistics is a very good predictor of population ancestry. I gave you a few pretty telling examples. Linguistics is not a turnkey solution, of course, and it takes time for it to build out genealogical classifications but, when it's in good hands, it becomes a reliable tool. In any case, it's good to have both - genetics and linguistics.

Onur said...

But mind you the magnitude of language replacement in Sub-Saharan Africa means that the foraging populations must have been rather sparse and not as linguistically diverse as their counterparts in the Americas or Papua New Guinea.

There is no such necessity. The restricted development of agriculture in the pre-1492 New World compared to the Old World can easily explain the differences between the past demographic proecesses of both sides of the world.

Eh? Genetic diversity is a function of population size. The invention of agriculture has brought about a cumulative increase in population size and the allelic diversity has gone up with it. Agriculture and pastoralism also resulted in population mixture on the continent-wide scale that also drove genetic diversity up. Haplotypic diversity and pairwise nucleotide differences are consistently higher in African agriculturalists than in African hunters and gatherers. Bantu "diverged" earlier than Pygmies. Take a look at Fig. 2 in "Phylogeography of the human mitochondrial L1c haplogroup: Genetic signatures of the prehistory of Central Africa" by Batini et al.

You are arguing against all the evidence. African hunter-gatherer populations have the high genetic diversity not only in Africa but also in the whole world.

No, that's very relevant. The post-1492 colonization of the New World is a normal population process and a good example of how genetic diversity has nothing to do with the age of a population. The New World is now more diverse than Sub-Saharan Africa because it sampled populations from every corner of the globe.

The post-1492 colonization of the New World is very remote from being a normal population process. It is unparalleled in the whole history of the Homo genus. So you are comparing apples and rocks.

German Dziebel said...

"You are arguing against all the evidence. African hunter-gatherer populations have the high genetic diversity not only in Africa but also in the whole world."

Onur, just open the paper I referenced and see for yourself. Batini compares Pygmies (HG) and Bantu (A) side by side and calculates TMRCA and other stats for the sublineages of L1 in the two groups. Everything that I say is always based on evidence.

"The post-1492 colonization of the New World is very remote from being a normal population process. It is unparalleled in the whole history of the Homo genus. So you are comparing apples and rocks."

And agricultural expansions were unparalleled in comparison with earlier foraging expansions. A 5-fold population increase. See here http://dienekes.blogspot.com/2011/04/global-demographic-expansions-followed.html

Genetic diversity just gets recycled between continents that serve like a "magnet" to human populations. After 1492 it's New World, before it was Sub-Saharan Africa, but it's part ad parcel of what human population history is all about. Great allelic diversity marks not a source but a destination for human populations.

Onur said...

On the individual level, yes, genetics can assign you to a unique cluster but on the group level linguistics is a very good predictor of population ancestry. I gave you a few pretty telling examples. Linguistics is not a turnkey solution, of course, and it takes time for it to build out genealogical classifications but, when it's in good hands, it becomes a reliable tool. In any case, it's good to have both - genetics and linguistics.

I can name you many counter examples, but I don't want to prolong that issue.

Two correction for my previous comment:

"proecesses of both sides of the world" -> "processes of..."

"African hunter-gatherer populations have the high genetic diversity not only in Africa but also in the whole world." -> "African hunter-gatherer populations have the highest genetic..."

Onur said...

German, I completely disagree with all of what you say about genetic diversity and repeat what I said: "it is against all the evidence". I hope you won't use Hadza, a ridiculously small population, as an example again.

German Dziebel said...

"I completely disagree with all of what you say about genetic diversity and repeat what I said: "it is against all the evidence".

You can repeat your opinion 50 times over, it still doesn't constitute evidence. Show me side by side comparison of agriculturalists, pastoralists and African foragers. I provided you with scientific, quantifiable evidence. You rejected it for the sake of a myth. And I'm not even talking about Hadza. We discussed Hadza earlier, and it's an important testimony to the original levels of genetic diversity in Africa, prior to population expansions and continent-wide admixtures. Batini compared Pygmies and Bantu. Check out another paper in which !San are less diverse in terms of pairiwise nucleotide differences than Nootka Amerindians despite having a larger demographic population size. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC52581/

Onur said...

German, we have discussed Hadza enough, so I don't want to open that discussion again (it is clear that we disagree in a lot of things there).

As for your request of evidence, I didn't present evidence because I didn't even see a need to present evidence for this countless times proven fact that the highest genetic diversity in humans is in African hunter-gatherers. Unlike you, I am now presenting a whole genome study as evidence, so will leave little room for doubt about this obvious fact:

http://www.nature.com/nature/journal/v463/n7283/full/nature08795.html#B1

Onur said...

Genetic diversity just gets recycled between continents that serve like a "magnet" to human populations. After 1492 it's New World, before it was Sub-Saharan Africa

Where is the evidence that Sub-Saharan Africa was ever a magnet for modern human populations rather than source?

German Dziebel said...

Onur, now that you pulled a paper, go to p. 22 Fig. 3 of its Supplementary Material and tell me what you're seeing. What I am seeing is that admixed South African Colored (SAC) and South African Europeans (SAE) have higher total and per autosome heterozygosity rates than the 5 indigenous South African individuals. ABT, which is a Xhosa/Sotho Bantu, is a close third. The 4 Khoisan individuals show altogether lower levels of heterozygosity than SAC, SAE and ABT.

Wake up, Onur, and study the data.

"(it is clear that we disagree in a lot of things there)."

What's there to disagree about? Hadza is an ancient, isolated, drifting, low-diversity, low-admixture population that has been drawn into pan-African gene flow relatively recently. Linguistically, it's the most divergent language in the Khoisan phylum.

Onur said...

Onur, now that you pulled a paper, go to p. 22 Fig. 3 of its Supplementary Material and tell me what you're seeing. What I am seeing is that admixed South African Colored (SAC) and South African Europeans (SAE) have higher total and per autosome heterozygosity rates than the 5 indigenous South African individuals. ABT, which is a Xhosa/Sotho Bantu, is a close third. The 4 Khoisan individuals show altogether lower levels of heterozygosity than SAC, SAE and ABT.

Wake up, Onur, and study the data.


German, the problem is that they made that heterozygosity analysis with SNP-based Illumina arrays, which are very far from capturing whole-genome variation in humans in its entirety and are highly biased towards Europeans in their SNP choice. They admit the unreliability of those heterozygosity results with these words in the full paper:

"The large number of novel SNPs raises concerns regarding the ability of current genotyping arrays to capture effectively the true extent of genetic diversity and haplotype structure represented in southern Africa. Assessing percentage heterozygosity for 1,105,569 autosomal SNPs using current-content Illumina arrays, we were surprised to find lower heterozygosity in KB1 compared to a region-matched European control (Supplementary Data and Supplementary Fig. 3a, b), because it is well known that genetic diversity is highest in Africa. However, analysis of whole-genome sequencing data for KB1 and ABT revealed high percentages of heterozygous SNPs (59% and 60%, respectively), as expected. This discrepancy underscores the inadequacy of current SNP arrays for analysing southern African populations."

What's there to disagree about? Hadza is an ancient, isolated, drifting, low-diversity, low-admixture population that has been drawn into pan-African gene flow relatively recently. Linguistically, it's the most divergent language in the Khoisan phylum.

I've discussed them with you enough, so won't discuss them further, at least for now.

German Dziebel said...

These are just excuses on the part of the authors of the paper. Sure, the SNP arrays are biased in favor of Europeans AND Bantu agriculturalists but against African foragers!

"Where is the evidence that Sub-Saharan Africa was ever a magnet for modern human populations rather than source?"

It's should rather be "where is the evidence that Sub-Saharan Africa was a ever a source for modern human populations than a magnet." Out of Africa theorists have never properly rejected the null hypothesis that higher intrapopulational genetic diversity is the result of population growth and admixture between formerly divergent and isolated populations characterized, instead, by higher interpopulation diversity values. In a slight-of-hand manner, they simply assumed that high intrapopulational genetic diversity is indicative of a homeland and its decrease with geography represents serial bottlenecks. It's like saying that everybody came from New York through a series of founder effects. We know that South Africa attracted a massive spread of agriculturalists from West Africa. We also know, with fewer details, that there was a pastoralist migration into South Africa from East Africa. For East Africa, the detailed argument that the signs of antiquity can easily be confused with signs of cumulative admixture, was made in a fairly recent paper blogged about by Dienekes. http://dienekes.blogspot.com/2009/08/mtdna-and-ethnic-differentiation-in.html
The Hadza example - admittedly singular - directly indicates that the expansions of foragers into South Africa, albeit not as massive as later agriculturalist expansions, resulted in elevated intrapopulational genetic diversity.

John Hawks also has a relevant post around the same theme. http://johnhawks.net/weblog/reviews/genomics/modern/henn-southern-africa-origins-2011.html. By "homogenization," BTW, he means the reduction of interpopulational diversity with the corresponding increase in the intrapopulational diversity.

Onur said...

Sure, the SNP arrays are biased in favor of Europeans AND Bantu agriculturalists but against African foragers!

Yes, also biased in favor of Bantu agriculturalists compared to African foragers, who are often neglected. There is nothing so strange about this as Bantu agriculturalists today constitute a very significant part of the population of Sub-Saharan Africa while African foragers are nearly on the brink of extinction. This is all to do with the commercial interests of the genetic testing companies.

Onur said...

German, also read Dienekes' refutation of me using the Denisovan paper in response to Errol's comment at the bottom of this thread:

http://dienekes.blogspot.com/2010/12/is-multi-regional-evolution-dead.html

So apparently African foragers are genetically much more divergent from the rest of modern humans than I had thought before.

Onur said...

By "homogenization," BTW, he means the reduction of interpopulational diversity with the corresponding increase in the intrapopulational diversity.

German, what you don't understand about agricultural expansions over forager territories is that they always reduced total genetic variation in places they occured. Because due to their sweeping nature they as a rule reduced inter-population diversity much more than they increased intra-population diversity. Look at for example how little genetic contribution Bantu agriculturalists got from the foragers they encountered in their expansion. So the genetic homogenizations resulting from agricultural expansions over forager territories are polar rather than balanced.

Dieneke, sorry for the triple posting; the discussion got too much diversified in its content.

German Dziebel said...

"German, also read Dienekes' refutation of me using the Denisovan paper in response to Errol's comment at the bottom of this thread:

http://dienekes.blogspot.com/2010/12/is-multi-regional-evolution-dead.html"

Onur, take a look at Table S6.2 referred to by Dienekes in his "clarification". You'll see that, while Han-Yoruba (.0811) divergence is smaller than San-Yoruba (.0913), Papuan-San divergence is much larger (.0933), while Han-San is very close to San-Yoruba (.0907). On the right hand side of the Table you'll see the divergence dates, wherein, again, Papuans and San diverged earlier then San and Yoruba.

The authors then make a misleading comment: "Table S6.2 also shows that the San are an “outgroup” with about equal divergence to the other present-day humans (Han, French, Papuan and Yoruba) at an average of 595,000 years." If you look at the Table, you'll see that, contrariwise, it's Papuans who stand out at 607, with San-Yoruba, San-Han and San-French showing similar values (593, 590, 589).

Not to mention the fact that Amerindians weren't even sampled, and Amerindians are usually showing greatest divergence from SSAfricans and especially the San.

"So apparently African foragers are genetically much more divergent from the rest of modern humans than I had thought before."

Not so if you read the data carefully. And when African foragers look more divergent it's probably because they have shifted away from Neanderthal-Denisovan heritage but haven't had enough time to get fixed in the new derived states.

"Yes, also biased in favor of Bantu agriculturalists compared to African foragers, who are often neglected. There is nothing so strange about this as Bantu agriculturalists today constitute a very significant part of the population of Sub-Saharan Africa while African foragers are nearly on the brink of extinction. This is all to do with the commercial interests of the genetic testing companies."

Using your logic, then, Amerindians would be the biggest victims of the ascertainment bias of all populations. Especially since many of them had gone extinct before any genetic companies came into being. The complete genome data seems to be pretty straightforward and consistent with common sense: as populations aggregate and grow in size, they gain diversity. Neanderthals apparently weren't diverse but they were much older than "African foragers." Their LD values and intergroup diversity must have been very high.

The authors simply encountered the results that went against the established consensus/myth and decided to question the data but not the consensus.

Onur said...

Once again, German, you are trying to fit the data to your pet theory and making strange interpretations meanwhile.

The authors then make a misleading comment: "Table S6.2 also shows that the San are an “outgroup” with about equal divergence to the other present-day humans (Han, French, Papuan and Yoruba) at an average of 595,000 years." If you look at the Table, you'll see that, contrariwise, it's Papuans who stand out at 607, with San-Yoruba, San-Han and San-French showing similar values (593, 590, 589).

There is nothing misleading about their comment, they are just stating what is obvious. San are clearly the outgroup among all modern humans according to the data and Papuan-San distance is a little bit more than the distance of San to the rest of modern humans most probably largely because of the Denisovan contribution in Papuans.

Not to mention the fact that Amerindians weren't even sampled, and Amerindians are usually showing greatest divergence from SSAfricans and especially the San.

I don't think sampling of Amerindians would change the outgroup situation of San.

Not so if you read the data carefully.

I am reading the data as unbiased and careful as possible, so it is you who is uncareful.

And when African foragers look more divergent it's probably because they have shifted away from Neanderthal-Denisovan heritage but haven't had enough time to get fixed in the new derived states.

This is pure speculation with no evidence.

Using your logic, then, Amerindians would be the biggest victims of the ascertainment bias of all populations. Especially since many of them had gone extinct before any genetic companies came into being.

There are hundreds of millions of Amerindian descendants (mostly mixed with other races) in the Americas, just go to Mexico.

The complete genome data seems to be pretty straightforward and consistent with common sense: as populations aggregate and grow in size, they gain diversity.

Read my previous post about agricultural expansions for an answer to this.

Neanderthals apparently weren't diverse but they were much older than "African foragers." Their LD values and intergroup diversity must have been very high.

We don't know about Neanderthal genetic diversity enough to make such judgements.

The authors simply encountered the results that went against the established consensus/myth and decided to question the data but not the consensus.

You have yet to refute the claims of the authors. Your comments are full of wishful thinking and unfettered speculation.

German Dziebel said...

"Because due to their sweeping nature they as a rule reduced inter-population diversity much more than they increased intra-population diversity. Look at for example how little genetic contribution Bantu agriculturalists got from the foragers they encountered in their expansion."

I can't talk about it in the abstract. I gave you an example of L1 in Pygmies and Bantu (the Batini paper), where agricultural sublineages are more diverse than forager sublineages. Gene flow between agriculturalists and foragers is indeed usually asymmetrical, with more forager women entering agricultural communities and bringing with them additional variation. It's the opposite for men.

But we don't need to restrict ourselves to the forager-agricultural dynamics. Look at Y-DNA hg E which spread all over Africa in pre-agricultural times. It does seem that in SS Africa, first, larger but younger groups of foragers first merged with smaller but older groups of foragers in the Pleistocene and then in the Holocene it's pastoralists and agriculturalists that expanded into the territories occupied by foragers plus established continent-wide patterns of gene flow across agricultural and pastoral communities. Overtime, genetic diversity has grown considerably among all SS Africans. Apparently, in East Asia and especially in Oceania and the Americas populations remained isolated from each other the longest.

Onur said...

German, just read Henn et al. 2011 and see how San are genetically the most diverse population worldwide, thus refuting your diverse agriculturalists thesis:

http://www.pnas.org/content/early/2011/03/01/1017511108.full.pdf+html

German Dziebel said...

"German, just read Henn et al. 2011 and see how San are genetically the most diverse population worldwide, thus refuting your diverse agriculturalists thesis."

Onur, it's you who should be reading papers. Henn et al. 2011 says (Suppl Mat, p. 2-3):

"Bantu-speaking populations, which have recently absorbed migrants during the Bantu expansion, have the highest haplotype heterozygosity at 100Kb windows; the hunter-gatherer Sandawe and ≠Khomani San groups have elevated haplotype heterozygosity, possibly in part due to recent gene flow from agriculturalist into these populations (Fig. 1)."

"Once again, German, you are trying to fit the data to your pet theory and making strange interpretations meanwhile."

I never do such kind of thing. Out of America is an objective possibility with just as much data behind it as out of Africa. I don't invent pet theories.

"I don't think sampling of Amerindians would change the outgroup situation of San."

You don't know what "outgroup" means. Amerindians (and sometimes Papuans) are just as good an outgroup as San. Amerindians tend to be the most divergent population from Africans.

"There are hundreds of millions of Amerindian descendants (mostly mixed with other races) in the Americas, just go to Mexico."

You picked it up from a recent paper, I understand, but it's irrelevant for the present discussion. The danger of an ascertainment bias in the Americas is the greatest because many regional Amerindians groups suffered a population collapse at the end of the 19th century and even now they prefer not to donate blood for samples. How many complete genomes from Amerindians are you aware of?

"This is pure speculation with no evidence."

Just look at the trees that have Denisovans and Neanderthals in them. Africans are the furthest removed from them, and these are true outgroups, as they obviously preceded humans.

"We don't know about Neanderthal genetic diversity enough to make such judgements."

We know about Neanderthal genetic diversity enough to appreciate its limited nature. Just look at blood groups: the 2 neanderthals sampled for classical markers ended up having blood group O, which, among modern human populations, is found at highest frequencies in the Americas. Alternatively we don't know anything about AMH and African archaic genes. In the absence of such direct information, all inferences of antiquity from diversity are spurious.

"You have yet to refute the claims of the authors. Your comments are full of wishful thinking and unfettered speculation."

The authors refuted themselves. And as time goes by geneticists produce more and more data that contradicts their own established "truths."

Onur said...

German, let's leave aside for the moment intra-Sub-Saharan population relationships and just focus on the relationships between Sub-Saharan Africa and the rest of the world. Sub-Saharan Africa, North Africa, Europe, West Asia, South Asia, East Asia and Southeast Asia are regions of the world long inhabited by crowded modern human populations. Of all these regions, Sub-Saharan Africa is unique with its extremely high genetic diversity/heterozygosity and extremely low linkage disequilibrium (Hadza are an exception, but their extremely low population size and relative isolation easily explains their difference from other Sub-Saharan populations) despite the fact that all these regions have had high population densities and intensive internal migrations for quite a long time. This uniqueness of Sub-Saharan Africa compared to the other regions I mention strengthen the view that Sub-Saharan Africa is the region inhabited by modern humans the earliest.

You don't know what "outgroup" means. Amerindians (and sometimes Papuans) are just as good an outgroup as San. Amerindians tend to be the most divergent population from Africans.

On a worldwide scale there is no better outgroup than Sub-Saharan hunter-gatherers, and they are followed by Sub-Saharan agriculturalists (including pastoralists).

We know about Neanderthal genetic diversity enough to appreciate its limited nature. Just look at blood groups: the 2 neanderthals sampled for classical markers ended up having blood group O, which, among modern human populations, is found at highest frequencies in the Americas.

You are talking about just two specimens. Is that your criterion of being enough in a complicated issue like genetic diversity?!

The authors refuted themselves. And as time goes by geneticists produce more and more data that contradicts their own established "truths."

I don't want to get polemical by responding to your this unsubstantiated and provocative claim.

BTW, in another Dienekes thread you write thus:

"Our time was up. Onur had been reading my string with Terry Toohill here on Dienekes, in which I referred to the anthropogenic theory of Pleistocene extinctions as a "conspiracy theory," and tossed it back at me."

http://dienekes.blogspot.com/2011/03/neandertal-genes-in-east-africa.html

My conspiracy theory analogy had absolutely nothing to do with your discussion with Terry Toohill; it is completely a figment of your imagination.

Onur said...

You picked it up from a recent paper, I understand, but it's irrelevant for the present discussion. The danger of an ascertainment bias in the Americas is the greatest because many regional Amerindians groups suffered a population collapse at the end of the 19th century and even now they prefer not to donate blood for samples. How many complete genomes from Amerindians are you aware of?

What recent paper are you talking about? I just referred to the very obvious fact that Amerindian descendants far outnumber known African hunter-gatherer descendants enough to catch the full attention of genetic testing companies for Amerindian genetics.

I never do such kind of thing. Out of America is an objective possibility with just as much data behind it as out of Africa. I don't invent pet theories.

Dienekes gave the best response to your pet theory.

ETHAN said...

I don't really see how high interpopulation diversity and low intrapopulation diversity in prehistoric Africa is inconsistent with the out of Africa hypothesis. Indeed, assuming that people tend to travel in groups, it seems like that would amplify the serial founder effect.

I was under the impression that genetic diversity among pre-Columbian Americans was low both within and between populations, although I admit I am not familiar with the research. If that is the case, where would the diversity come from in an out of America scenario? All the examples you give of destinations having high diversity rely on admixture of dissimilar groups.

I am also inclined to believe genetic evidence over linguistic in this case, because language replacement that preserves genetic diversity is quite plausible and historically well documented, while it is hard to imagine total population replacement preserving linguistic diversity.

Bayu Ardiyanto said...

I was very impressed, this is a very interesting post,, thank you for sharing