September 01, 2010

Three-stage expansion of humans across Eurasia and into the Americas

This is a fairly good paper. Many articles in the literature either date human expansions by genetic methods (which tell us about the age of common ancestors, or the accumulated genetic variation and its characteristics -e.g., linkage disequilibrium- but tell us little about where the ancestors lived), or by geo-demographic methods (which tell us about how populations grow and expand on the map given various parameter settings), or by archaeological methods (which record the spatio-temporal occurrence of sites, but are difficult to interpret and are subject to various biases).

The current paper combines the latter two methods to present a picture of the spread of humans across northeastern Eurasia (from southern Siberia) and eventually into the Americas. The point of origin (southern Siberia) is inferred via the dating of earliest archaeological sites, population growth and expansion is modeled via a simple diffusion equation, which is reality-checked via calibrations at other sites.

The authors make a good point about the mess that population genetics is with respect to dating human movements, so their decision to avoid it is understandable. However, genetics will eventually be able to complement the other two approaches, by the study of ancient DNA from the different sites (whenever they are associated with human remains). So far, ancient DNA in the region has been limited to Holocene remains, which postdate the events of interest, but if 30,000 year old Kostenki can yield useful results, I don't see why human remains from northeastern Eurasia cannot, as conditions for DNA preservation there (=cold) are good.

While archaeological dates (which are based on physics and have small confidence intervals) are far more secure than genetic ones, they have the problem of "who done it", as we can never be sure that the occupants of an early site are the same people as the later occupants of the same site. This can be done -to an extent- with craniometry, the fourth method of estimation human movements, but (a) interpreting population continuity in the face of ongoing adaptation of the human neurocranium is difficult and (b) you can get DNA from very incomplete bone and tooth remains, but you need a mostly complete skull to make statistical use of it using craniometry.

There have been long debates about the identity of the Paleondians, the long-skulled early inhabitants of the Americas, that do not resemble modern Amerindians, but DNA analysis of several ancient examples from both north and south America has revealed similar types of DNA as those found in the current inhabitants.

PLoS ONE 5(8): e12472. doi:10.1371/journal.pone.0012472

Archaeological Support for the Three-Stage Expansion of Modern Humans across Northeastern Eurasia and into the Americas

Marcus J. Hamilton, Briggs Buchanan

Abstract

Background
Understanding the dynamics of the human range expansion across northeastern Eurasia during the late Pleistocene is central to establishing empirical temporal constraints on the colonization of the Americas [1]. Opinions vary widely on how and when the Americas were colonized, with advocates supporting either a pre-[2] or post-[1], [3], [4], [5], [6] last glacial maximum (LGM) colonization, via either a land bridge across Beringia [3], [4], [5], a sea-faring Pacific Rim coastal route [1], [3], a trans-Arctic route [4], or a trans-Atlantic oceanic route [5]. Here we analyze a large sample of radiocarbon dates from the northeast Eurasian Upper Paleolithic to identify the origin of this expansion, and estimate the velocity of colonization wave as it moved across northern Eurasia and into the Americas.

Methodology/Principal Findings
We use diffusion models [6], [7] to quantify these dynamics. Our results show the expansion originated in the Altai region of southern Siberia ~46kBP , and from there expanded across northern Eurasia at an average velocity of 0.16 km per year. However, the movement of the colonizing wave was not continuous but underwent three distinct phases: 1) an initial expansion from 47-32k calBP; 2) a hiatus from ~32-16k calBP, and 3) a second expansion after the LGM ~16k calBP. These results provide archaeological support for the recently proposed three-stage model of the colonization of the Americas [8], [9]. Our results falsify the hypothesis of a pre-LGM terrestrial colonization of the Americas and we discuss the importance of these empirical results in the light of alternative models.

Conclusions/Significance
Our results demonstrate that the radiocarbon record of Upper Paleolithic northeastern Eurasia supports a post-LGM terrestrial colonization of the Americas falsifying the proposed pre-LGM terrestrial colonization of the Americas. We show that this expansion was not a simple process, but proceeded in three phases, consistent with genetic data, largely in response to the variable climatic conditions of late Pleistocene northeast Eurasia. Further, the constraints imposed by the spatiotemporal gradient in the empirical radiocarbon record across this entire region suggests that North America cannot have been colonized much before the existing Clovis radiocarbon record suggests.

Link

55 comments:

German Dziebel said...

This paper is also being discussed on Razib's blog:

http://blogs.discovermagazine.com/gnxp/2010/08/the-new-world-in-three-easy-steps/#comments

Robert said...

Spencer Wells said on PBS that as little as 10 to 20 people are ancestral for all the people of the Americas.

Spy said...

Speaking of craniometry, Dienekes—you still have your Anthropometric Calculator up. But that thing returns "Greek" for everybody. Just play with it!

eurologist said...

Figures 4A and B are relatively uncontroversial, as most researchers would agree that the earliest occupations occur in southern Siberia, most likely from central Asian populations between ~50k-45k BP,...

While I agree with this, I am still struggling with that path from "central Asia." IMO the easiest path would have been via Pakistan, Afghanistan, and along the west side of the Himalayas. And indeed, this seems a likely path for the eventual population of Europe at almost exactly the same time.

However, genetic evidence points to different, eastern source population, east of the Himalayas. As I have commented before, I think that groups originating from the Brahmaputra river region could have arrived in Siberia via northern Myanmar and adjacent west China as quickly, if not faster than via a coastal SE Asian route.

Eventually, it should be possible to clarify this via genetic evidence...

terryt said...

"Our results show the expansion originated in the Altai region of southern Siberia ~46kBP"

That puts early humans way north of where anyone I've argued with would be prepared to accept. Maju for example has argued that the evidence for more than 60,000 year old human presence right across Central Asia from the Altai to the Upper Amur Basin is some pre-modern population. I've always believed that humans had made it quite some way north early in the piece.

"I am still struggling with that path from 'central Asia'".

Presumably exactly the route I just mentioned.

"IMO the easiest path would have been via Pakistan, Afghanistan, and along the west side of the Himalayas".

Perhaps the easiest in an east/west direction, but extremely difficult from south to north.

"I think that groups originating from the Brahmaputra river region could have arrived in Siberia via northern Myanmar and adjacent west China as quickly"

But the movement of Tibeto-Burman speakers convincingly argues for movement from north to south in that region. Perhaps earlier movements were in the other direction, but we have no evidence for that. Just postulation.

"I think that groups originating from the Brahmaputra river region could have arrived in Siberia via northern Myanmar and adjacent west China as quickly, if not faster than via a coastal SE Asian route".

But neither route would be anything like as fast as a route through the south-facing grassland regions of Central Asia.

eurologist said...

But neither route would be anything like as fast as a route through the south-facing grassland regions of Central Asia.

Can you be more specific? I would like to know more exactly where your proposed source population is, and the exact path, so I can better understand what you mean.

Getting to Altai, and getting there first (although we have known for a long time) is just mind-boggling.

Getting there from Afghanistan and Kyrgyzstan ( place that most people believe were steeping stones on the way to Europe) is relatively straightforward, but does not fit the genetic picture (at least not a parsimonious one: you would need to postulate two distinct migration events, annihilation of one of the source populations, and more).

Getting there the other way-around than migrations progressed (i.e., from the NE coast) is also relatively easy, but completely contradicts the archaeological finds.

Getting there from the south means carefully threading frigid and barren deserts and mountain passes, with only a few regions in between that could have developed sufficiently large buffer populations in the short time most envision.

One such possible path is Chengdu --> Lanzhou --> Zhangye --> Altai Mountains --> Altai. You almost needed a satellite map for that one, even 50,000 years ago ;)

eurologist said...

Also, to clarify, I know we recently discussed brief northward excursions into the Tibetan Plateau during suitable prehistoric times. However, my understanding was that the plateau was not warm and wet enough to sustain abundant animal life at around 50,000 years ago - even though many adjacent arid regions for a brief interval during that time did support lush grasslands.

Otherwise, of course you could reach the Altai region simply by going north from Lhasa...

Andrew Oh-Willeke said...

I'm a bit surprised that a trans-Atlantic route considered legitimate enough to win any mention in the abstract.

The real story in the paper seems to be the sort of thing that could be summed up in a single map, or perhaps a slide show of sequential maps giving archeological dates.

As a Devil's Advocate, in fairness to the argument contrary to the one that "North America cannot have been colonized much before the existing Clovis radiocarbon record suggests," the argument for earlier dates seems to be that there were a handful of small civilizations in coastal areas which much older dates allegedly supported by radiocarbon dating. There is great dispute over the methodology that backs up the very old radiocarbon dates. But, if the old dates are right, no diffusion model can falsify them. An older date would imply there there was a small early wave of cherry picking coastal colonists that established a few communities but didn't expand much (perhaps they didn't know how to hunt terrestrial animals effectively, or failed to adapt to new fish and plant resources very well).

The genetic evidence seems suggestive of the idea that the early Native Americans formative community may have had predominant ancestry from East Asia, with a minority contribution from Central Asia, with there being some population structure distinguishing the two source populations even after arrival in North America. So, a claim that "the expansion originated in the Altai region of southern Siberia ~46kBP" seems to obscure some of the likely pre-history suggested by the genetic evidence. Genetic evidence isn't great at telling you when waves of migration happened (at a level of detail accurate enough to be interesting), but it does shed some light on how many waves of migration one should expect to see and the deep geographic origins, at a general level, of each wave.

terryt said...

"Can you be more specific? I would like to know more exactly where your proposed source population is, and the exact path, so I can better understand what you mean".

I did find a series of articles while I was arguing with Maju over a Central Asian route east. I'll see if I can find them again when I have time. The main point made in the articles was that humans had been present in the hilly region that stretches across north of Mongolia, from the Altai to the Upper Amur River, virtually continuously for about 160,000 years. Maju's counter was that they were probably not 'modern' humans. However I suggest that modern humans mixed with the population and then spread through the region.

"I know we recently discussed brief northward excursions into the Tibetan Plateau during suitable prehistoric times".

So my guess is the route had nothing to do with the Tibetan Platea. Modern humans probably moved north to the Altai via the Kazakh steppe. It seems that mammoths and other megafauna lived through much of the region at the time, so there must have been enough suitable habitat for them right across to northern China. In fact the ~46kBP date supports the idea that humans were responsible for the extinction of the steppe mammoth, long before they had even entered the woolly mammoth's habitat.

"Getting there from Afghanistan and Kyrgyzstan ( place that most people believe were steeping stones on the way to Europe) is relatively straightforward, but does not fit the genetic picture (at least not a parsimonious one: you would need to postulate two distinct migration events, annihilation of one of the source populations, and more)".

Perhaps. But a case can easily be made for it. Y-hap C is found right across the region, from C5 in Pakistan, northern India Kashmir and Nepal, to C3 in Mongolia and northern China. Of course most prefer to see C3's origin in Eastern Mongolia and to link Y-hap C with a southern route through India. But we have C1 in Japan, C2 in Wallacea and C4 in Australia. The big gap in C's distribution is through SE Asia and eastern India. And one would expect 'annihilation of one of the source populations' (well groups of them anyway) across Central Asia as the climate changed periodically.

It is also quite possible to argue a similar pattern for some mtDNA N haplogroups.

German Dziebel said...

"Y-hap C is found right across the region, from C5 in Pakistan, northern India Kashmir and Nepal, to C3 in Mongolia and northern China. Of course most prefer to see C3's origin in Eastern Mongolia and to link Y-hap C with a southern route through India. But we have C1 in Japan, C2 in Wallacea and C4 in Australia. The big gap in C's distribution is through SE Asia and eastern India."

This gap seems to be filled by C's phylogenetic cousin, namely hg D, though, which is geographically nested between the extremes of hg C and converges with C in Japan (Ainu) but isn't found in the Americas. However, hg D's sister haplogroup E is found outside of the range of hg C, namely all over Sub-Saharan Africa (and Europe). Hence, the annihilation event is unlikely.

terryt said...

"This gap seems to be filled by C's phylogenetic cousin, namely hg D, though"

Haplogroup D hardly fills the gap. It's certainly very much present in the Andamans and in Tibet, but is very sparse anywhere else. As far as I'm aware it's not found in India or very far into SE Asia. And D is no more 'C's phylogenetic cousin' than are any of the F-derived Y-haps.

Michele said...

Since Levallois lithics have been found and described by Dr. Barbara Purdy from pre-Clovis levels in northern Florida, and the oldest tools have been dated to 26-28,ooo years old by both thermoluminescence and patination studies,(Purdy 1981,2008), this argument would seem to be a moot point.Mark Corbitt

German Dziebel said...

Mark: check out comments on Razib's website. The early dates at CCA are important as direct disproval of an assumption such as Nettle's that Pleistocene archaeology is a static field against which we should be benchmarking the age of Amerindian linguistic diversity.

Boasian said...

The suggested presence in Northeast Asia of founding or nodal mitochondrial DNA haplotypes for each Amerindian haplogroup (Torroni et al. 1993a) precludes an alternative explanation, that of admixture between Northeast Asians and Amerindians. This “reverse migration” out of the Americas provides a backdrop for the later formation of Sea Mammal Hunting Cultures, an idea Franz Boas identified as "Eskimo wedge theory" (Boas 1905 and 1910; also see Ousley and others in, Human Biology, June 1995). Presumably, as Boas believed, the removal of glacial barriers allowed human contact between the Longitudinal Hemispheres that was geographically un-encumbered and evidenced by widespread Holocene acculturation (also see Ackerman 1982; Dumond 1983; and Heizer 1943; and others) inasmuch as earlier Amerindian tribal populations existed before deglaciation. These assumptions provide an alternative explanation to mtDNA analyses, challenging the idea that Amerindian mtDNA Lineages (AAM1, CAM43, and DAM88) found in Siberians are ancestrally linked to the initial colonizers of the Americas. Rather, mtDNA analysis could be seen to support the Boas data in that the formation of contemporary Circumarctic Populations in Siberia may have been influenced by post-glacial Amerindian movements into Beringia, Siberia, and Northeast Asia.

onur said...
This comment has been removed by the author.
onur said...

German, even if one day it is archaeologically proven that the ancestors of Amerindians entered the Americas much earlier than current mainstream dates (for instance, 30,000, 40,000 or even as early as 50,000 years BP), I don't think that the Out of Africa theory will ever be replaced by the Out of America theory as the standard model of anatomically modern human (AMH) origins. That the Sub-Saharan Africans are more ancient than all other AMH populations is one of the most solid facts of AMH genetics (note that I said genetics, not archaeology, as you disdain archaeology). I am sure none of your favorite scholars support the Out of America theory, which I only heard from you.

German Dziebel said...

Onur: you speculate, you predict, you ventriloquise, you demand proof, you absolve yourself from proving... And it's all pointless because you can't focus on the simple nature of my argument. I don't disdain archaeology. Every discipline should have its place and stick to the answers that it inherently can provide. Archaeology is a poor predictor of origins because artifacts are not heritable units. It's been misused for quite some time, and unfortunately it has misguided genetics by supposedly providing secure dates for the peopling of the Americas.

"I am sure none of your favorite scholars support the Out of America theory, which I only heard from you."

Why is this a problem? What is important at this point is that all mainstream linguists rejected all the attempts by geneticists and other "hard" scientists to pull linguistic data into an interdisciplinary argument for the out-of-Africa and recently-into-the-America. Geneticists tend to resort to such outcasts as Greenberg and Ruhlen to correlate their phylogenies with linguistic classifications. Also, when geneticists became interested in African click languages as supposedly a linguistic correlate of the earliest genetic haplogroups, mainstream linguists responded by saying that there's nothing about click languages that suggests great antiquity. When geneticists proposed that African Pygmies are supposedly a whopping 70,000 years old population, linguists only shrugged their shoulders, as Pygmies speak "normal" Niger-Congo languages and the evidence for a substratum is very tenuous.

"the standard model of anatomically modern human (AMH) origins..."

Be mindful of the fact that all extant human populations are not only anatomically but also behaviorally modern humans. These two things are tightly intertwined in all human societies. Behavioral modernity is attested in Africa and Europe at no earlier than 45,000 YBP. We know that AMH were replaced by Neandertals in West Asia. Hence, it's time to show some proof that AMH are in fact ancestral to modern human populations and that anatomical modernity predates behavioral modernity by tens of thousand of years. It's quite possible that AMH were replaced by Neandertals and by our true ancestors, namely behaviorally modern humans (who were also, of course, anatomically our ancestors). The way to show that AMH in Africa are our ancestors is to dig up some DNA in Africa and identify mtDNA L lineages in AMH remains. So far, no L lineages have ever been detected in any ancient remains of behaviorally modern humans in or out of Africa. Even Taforalt remains in North Africa (12K) are devoid of L lineages.

onur said...
This comment has been removed by the author.
onur said...
This comment has been removed by the author.
onur said...

German, you accuse of of not presenting proofs, but most of your proofs aren't proof at all, but your own interpretation into the data often in a distortive way. So I will go step by step to show you this.

And it's all pointless because you can't focus on the simple nature of my argument. I don't disdain archaeology.

I was being sarcastic when I made that remark. But don't forget that behind every sarcasm is a partial truth.

Every discipline should have its place and stick to the answers that it inherently can provide.

Yes, hence it is wrong to place linguistics at a higher position than genetics and archaeology when dealing with human origins and dispersals.

Archaeology is a poor predictor of origins because artifacts are not heritable units.

But it is surely a much better predictor than linguistics over tens of thousands years.

It's been misused for quite some time, and unfortunately it has misguided genetics by supposedly providing secure dates for the peopling of the Americas.

Misguided according to your linguistic interpretation.

What is important at this point is that all mainstream linguists rejected all the attempts by geneticists and other "hard" scientists to pull linguistic data into an interdisciplinary argument for the out-of-Africa and recently-into-the-America. Geneticists tend to resort to such outcasts as Greenberg and Ruhlen to correlate their phylogenies with linguistic classifications.

Surely geneticists arrive at the same conclusions with or without Greenberg's thesis.

Also, when geneticists became interested in African click languages as supposedly a linguistic correlate of the earliest genetic haplogroups, mainstream linguists responded by saying that there's nothing about click languages that suggests great antiquity.

Indeed, there is nothing about click languages that suggests great antiquity, but also there is nothing about click languages that doesn't suggest great antiquity as there is no way of knowing whether clicks are very ancient or not, at least linguistics doesn't give answers to such a question.

continued

onur said...

When geneticists proposed that African Pygmies are supposedly a whopping 70,000 years old population, linguists only shrugged their shoulders, as Pygmies speak "normal" Niger-Congo languages and the evidence for a substratum is very tenuous.

How is it possible to know whether current Pygmy languages have substratum or not without knowing anything about their earlier languages?! Also there is no fixed rules about substratum effects of earlier languages, people can switch to another languages without leaving behind any clear trace from the previous language (perhaps because the traces are too subtle). Anyway, according to Wiki, there may actually be substratum in current Pygmy languages:

"Some 30% of the Aka language is not Bantu, and a similar percentage of the Baka language is not Ubangian. Much of this vocabulary is botanical, deals with honey collecting, or is otherwise specialized for the forest and is shared between the two western Pygmy groups. It has been proposed that this is the remnant of an independent western Pygmy (Mbenga or "Baaka") language."

http://en.wikipedia.org/wiki/Pygmy_peoples#Ancestral_relationship

Behavioral modernity is attested in Africa and Europe at no earlier than 45,000 YBP.

It might be much older than that in Africa according to the recent discoveries there like Blombos Cave in South Africa.

It's quite possible that AMH were replaced by Neandertals and by our true ancestors, namely behaviorally modern humans (who were also, of course, anatomically our ancestors).

You are accusing me of speculation while you basically offer nothing but speculation and twisting of truth (I am speaking in general).

So far, no L lineages have ever been detected in any ancient remains of behaviorally modern humans in or out of Africa.

Can you be more specific? You are speaking as if there are many ancient DNA samples from Sub-Saharan Africa.

Even Taforalt remains in North Africa (12K) are devoid of L lineages.

That is ridiculous. Taforalt is north of Morocco, of course they will have Eurasian ilineages instead of L lineages. During the last glacial period the Sahara Desert was bigger than its current extent and more inaccessible, what else would you expect?

onur said...

Dieneke,

Sorry for the deletions. Because of its big size, the first part of my post was deleted out of my control after the posting of the second part, so I had to delete the second part.

onur said...

Typo: "people can switch to another languages"

It should have read:

"people can switch to another language"

onur said...

Typo: "Eurasian ilineages instead of L lineages"

"Eurasian lineages instead of L lineages"

Dienekes said...

onur, I told you not to triple post, and I see that you 7-posted. This is unacceptable behavior. I already told you to check your posts before submitting them. If you make a mistake, that is no reason to inundate the blog with useless comments. If one of your comments is too big, you can always split it into two. If you need to split it into 3 or more, then what you should do is edit your comment to make it smaller rather than flood the blog with comments, and then add even more to tell me you're sorry about it.

This is the very last warning.

onur said...

Dieneke, I know it is meaningless to say sorry at this point, but I apologize anyway as I hate to cause such technical problems in such a serious and well-prepared blog.

German Dziebel said...

As Dienekes hasn't yet posted my original reply to Onur, I had time to sharpen it a little bit. Please use this version.

"Yes, hence it is wrong to place linguistics at a higher position than genetics and archaeology..."

What is wrong is to dismiss linguistics, kinship studies and other "cultural" data as less accurate than archaeology or genetics. And that's what you do. I'm saying that every discipline should benchmark its results against the results and the methodology of other disciplines.

"But it is surely a much better predictor than linguistics over tens of thousands years."

Using the work "surely" doesn't a proof make. Over the past hundred years archaeology hasn't furnished any technological complex in Siberia or Alaska that can be considered ancestral to Clovis in America. Meanwhile, so many people are convinced that America was peopled from Siberia around the time Clovis emerges in the archaeological record. This is just an example of how archaeologists fake prehistoric expertise without providing a very simple illustration for their beliefs.

"Misguided according to your linguistic interpretation."

No. Even without linguistics, there's no evidence for the origin of Clovis.

"Surely geneticists arrive at the same conclusions with or without Greenberg's thesis."

What thesis? Geneticists disproved Greenberg's interpretation of non-linguistic evidence for a tripartite New World classification idea multiple times throughout the 1990s. On the other hand, it's interesting that geneticists felt compelled to tie their global phylogenies to linguistic classifications. Because, unlike you, scientists look for consistency across disciplines. Also because archaeological data, unlike genetic and linguistic data, is a poor material for phylogenies.

"Indeed, there is nothing about click languages that suggests great antiquity..."

There's an element of uncertainty in all interpretations - whether genetic or linguistic. But it were geneticists who tried to bring clicks to bear on the out of Africa model of human dispersals. Again, you see that geneticists desire an alliance with linguistics because there's nothing else to map their data onto. Again, archaeology is just not the discipline to study origins. What is also noteworthy is that geneticists resort to cultural stereotypes – if an African language has strange sounds, they must be “primitive” and more “ape-like” than other sounds.

What's interesting about clicks is that they are African-specific. Just like geneticists' African-specific haplogroups (Y-DNA E, A, B, or mtDNA L). And a phenomenon that's specific to a single geographic area is likely to be of derived nature. That's one of the reasons why linguists don't see clicks as the "original human sound."

German Dziebel said...

(contd.)

"Anyway, according to Wiki, there may actually be substratum in current Pygmy languages..."

This is Bahuchet's idea. He hasn't published much about it. And it's not confirmed by other linguists. Even if it turns out that floral and faunal vocabulary in some Pygmy languages has no parallels in Niger-Congo languages, it can also be explained as the existence of specialized vocabulary in foraging communities that didn't survive in agricultural speech communities.

"It might be much older than that in Africa according to the recent discoveries there like Blombos Cave in South Africa."

The sporadic signs of symbolic behavior in Africa only attest to the fact that modernization was taking place independently in several hominin lineages. At 45-40K years, everywhere in the Old World, we have in the archaeological record a systematic revolution taking place that continues uninterrupted into the present. This is the signature of our behavior. And it must have occurred in conjunction with linguistic diversification. The earliest signal for this diversification, however, comes not from Africa but from America.

"Can you be more specific? You are speaking as if there are many ancient DNA samples from Sub-Saharan Africa."

There're none. My point is science gotta have those DNA samples as a pre-requisite to marketing out of Africa to the public as "absolute truth." So far it's just a hypothesis. I gave you the example of Taforalt as the only available African DNA sample. Taforalt mtDNA showed no L lineages but they did have U6 lineages, which is a signature of Upper Paleolithic backflow to Africa. So we have evidence for an ancioent migration into Africa but none of a migration out of Africa. L lineages aren't found in ancient remains in Europe or Asia (comp. Kostenki is U2). L lineages aren't found at low frequencies in any populations along the putative southern route of expansion out of Africa (Andaman islanders are Y-DNA D and mtDNA M). If there was an expansion out of Africa (read: African genetic continuity outside of Africa) then we would expect to find typical African lineages plus derived non-African lineages outside of Africa. We don't have those. So, across ancient and extant DNA out-of-Africa signatures are poorly attested.

onur said...

What is wrong is to dismiss linguistics, kinship studies and other "cultural" data as less accurate than archaeology or genetics. And that's what you do. I'm saying that every discipline should benchmark its results against the results and the methodology of other disciplines.

Historical (including prehistory) linguistics is a too speculative discipline (I say discipline as I don't think it is a science). Rather than empirical research, it is primarily based upon speculations (it couldn't be otherwise given the extreme sparseness of relevant data). So why should I take it seriously when dealing with so ancient times like >10K BP?

Over the past hundred years archaeology hasn't furnished any technological complex in Siberia or Alaska that can be considered ancestral to Clovis in America.

Even if you are right on that point, that doesn't affect my thoughts as I am not a supporter of the Clovis First hypothesis (I am open to earlier dates for human arrival into the Americas). On the other hand, you have to acknowledge that archaeological evidence is usually fragmentary for such ancient times, especially in so sparsely-populated places like the Arctic and Subarctic regions, so cultural continuities (and discontinuities) are very hard to be detected.

What thesis? Geneticists disproved Greenberg's interpretation of non-linguistic evidence for a tripartite New World classification idea multiple times throughout the 1990s.

Certainly I am not a supporter of Greenberg's hypothesis. What I meant in my above post was that Greenberg's hypothesis doesn't affect the results of genetic studies, irrespective of whether they are consistent with Greenberg's hypothesis or not.

On the other hand, it's interesting that geneticists felt compelled to tie their global phylogenies to linguistic classifications. Because, unlike you, scientists look for consistency across disciplines. Also because archaeological data, unlike genetic and linguistic data, is a poor material for phylogenies.

I already wrote what I think about historical linguistics above.

There's an element of uncertainty in all interpretations - whether genetic or linguistic... That's one of the reasons why linguists don't see clicks as the "original human sound."

Whether the original human language had clicks or not is a very speculative subject, so don't expect me to seriously discuss it.

onur said...

Even if it turns out that floral and faunal vocabulary in some Pygmy languages has no parallels in Niger-Congo languages, it can also be explained as the existence of specialized vocabulary in foraging communities that didn't survive in agricultural speech communities

Yes, this is also possible. But if the 30% figure is true, I assign your idea a lower probability as 30% is too much for not being able to trace some proportion of them to the known languages or language families.

The sporadic signs of symbolic behavior in Africa only attest to the fact that modernization was taking place independently in several hominin lineages.

Climatic conditions in Sub-Saharan isn't suitable for long term preservation of archaeological artifacts. So this may at least partially explain the sporadic nature of the early evidence for behavioral modernity.

At 45-40K years, everywhere in the Old World, we have in the archaeological record a systematic revolution taking place that continues uninterrupted into the present.

But overwhelming majority of that archaeological record is from outside Sub-Saharan Africa. This isn't surprising as climatic conditions are more suitable for long term preservation of archaeological artifacts in much of the world outside Sub-Saharan Africa and as 45-40K years BP correspond to the times when modern human populations were gaining relatively high population numbers outside Sub-Saharan Africa for the first time.

The earliest signal for this diversification, however, comes not from Africa but from America.

Lots of speculations again.

Taforalt mtDNA showed no L lineages but they did have U6 lineages, which is a signature of Upper Paleolithic backflow to Africa. So we have evidence for an ancioent migration into Africa but none of a migration out of Africa.

Can you make the same statement for Sub-Saharan Africa instead of the vague word "Africa"? I am asking you this because North Africa isn't Sub-Saharan Africa, it has been much much more a part of the West Asian ecological zone (including humans) than the Sub-Saharan African one since the appearance of the Sahara Desert.

L lineages aren't found in ancient remains in Europe or Asia... So, across ancient and extant DNA out-of-Africa signatures are poorly attested.

Genetic phylogenies tell us that all non-African haplogroups are ultimately derived from just a relatively small subset of African haplogroups.

German Dziebel said...

"Genetic phylogenies tell us that all non-African haplogroups are ultimately derived from just a relatively small subset of African haplogroups."

"Historical (including prehistory) linguistics is a too speculative discipline (I say discipline as I don't think it is a science)."

These two statements are good indicators that you simply believe in the "powers" of genetic evidence and berate linguistics as a substandard pursuit. This an absolutely unprofessional attitude. Linguistics as a science is older than genetics, has larger databases (mine alone, for instance, encompasses kinship vocabularies from 2500 languages) and describes our critical differentiator from apes and hominins. Geneticists even routinely use languages and language families in their analyses (e.g., they say "Bantu" and not geographically defined population A). Not to mention the curious historical fact that Darwin was fascinated with linguistic phylogenies (so nascent in those days) as the way to arrive at subspecific classifications.

As for genetic phylogenies, they are unproblematic only on paper (or computer screen). Phylogeographically, their ability to support the out of Africa model borders on absurdity. For instance, mtDNA macrohaplogroup N is not indigenous in Africa, it expands in East Asia and shows up in Africa only in the form of derived clades such as N1 and U6. So we have a huge geographic gap between a putative African ancestor and the first non-African descendants. mtDNA macrohaplogroup M shows virtually the same pattern: it's not indigenous in Africa, it expands in South/East Asia and it's attested in Africa only in the form of a derived clade, namely M1. mTDNA L6 haplogroup, which is the most divergent lineage within the L3'4'5'6 clade is attested only outside of Africa, namely in West Asia. Y-DNA C haplogroup, which covers a huge geographic terrain from Australia to North America, is not attested in Africa. Y-DNA E haplogroup that accounts of more than half of Sub-Saharan African chromosomes is nested within a non-African CDEF clade and is directly related to only haplogroup D, which is a non-African haplogroup. So, what phylogeography tells us is that there're non-African clades (such as mtDNA M and N, or Y-DNA C, D, F) and there're African clades (mtDNA L0, L1, L2, Y-DNA A and B). How they are related to each other is unclear. But what we know in the very least is that the non-African clades moved into SS Africa at 40-45K and possibly admixed with the pre-existing African lineages. Unless those pre-existing African lineages are simply misclassified descendants of those incoming non-African lineages.

onur said...

German, I was talking about historical linguistics, which sometimes makes bold claims about language histories. I think historical linguistics works to some extent for the last several thousand years at most, but becomes useless beyond that as at that point assumptions and unknowns become too excessive making it practically impossible to make scientifically valid inferences.

Regarding haplogroup phylogenies, I will only say that the last two decades of research has only strengthened the scientific validity of the general Africa-rooted phylogenetic framework, not weakened it. BTW, L6 also exists in East Africa and probably originated there.

German Dziebel said...

"BTW, L6 also exists in East Africa and probably originated there."

Yes, it has been found in Egypt and Ethiopia. But it's lower frequency and lower diversity in Africa suggests that it could've come from Yemen or Oman in the course of the migration of Semitic-speakers into Africa.

"I will only say that the last two decades of research has only strengthened the scientific validity of the general Africa-rooted phylogenetic framework."

I just criticized it using the facts that have come to light in the last 10 years.

"which sometimes makes bold claims about language histories."

I hear you. Just stay away from long range comparison a la Greenberg, which creates artificial groupings such as Amerind and then dates them using shaky dates derived from archaeology.

"assumptions and unknowns become too excessive making it practically impossible to make scientifically valid inferences."

I don't think these assumptions and unknowns are greater in linguistics than in archaeology or genetics. It's precisely because these assumptions and unknowns are real that we have to compare what different disciplines have to say, without falling in love with one or two disciplines. In kinship studies, for instance, the patterns of evolutionary transformations within kinship vocabularies (see "My Genius of Kinship") have been tested against large samples and within many well-established language families. And it's been known for the past 150 years that African kinship systems are lacking in all the ancestral patterns attested in parts of Asia, Australia, Oceania and, most importantly, America.

terryt said...

"Yes, it has been found in Egypt and Ethiopia. But it's lower frequency and lower diversity in Africa suggests that it could've come from Yemen or Oman in the course of the migration of Semitic-speakers into Africa".

Possibly. But it's greater diversity in Yemen and Oman is quite likely to be product of its arrival there from both of the regions it is found in Africa: Egypt and Ethiopia. That would also explain its apparent lower diversity in Africa. Is its diversity lower if we include both regions?

"African kinship systems are lacking in all the ancestral patterns attested in parts of Asia, Australia, Oceania and, most importantly, America".

Interestingly those regions are basically marginal to the Eurasian human habitat. So it's quite possible that ancient kinship systems have been swept aside through Africa and much of Eurasia with the expansion of more recent systems.

German Dziebel said...

"Interestingly those regions are basically marginal to the Eurasian human habitat. So it's quite possible that ancient kinship systems have been swept aside through Africa and much of Eurasia with the expansion of more recent systems."

Why? On what grounds, Terry? I just debunked this stereotype of massive Old World language and kinship extinctions on Razib's site. African and European kinship systems were transformed away from the ancestral forms precisely because these areas were colonized from Asia. Even Khoisan systems have identifiable antecedents in Asia, Australia and America, and the Khoisans are the clear outliers in Africa.

Also, if there was no strong gene flow between an original departing population and a population that stayed behind isolation must be the strongest for the autochthonous population. If humans departed from Africa, we would have seen the survival of ancestral kinship forms in Africa and their stepwise transformation outside of Africa. Increasing population size is one of the main drivers behind the changes in kinship systems. As human populations grew in size, under the pressure to colonize the wide expanses of Eurasia and Africa, their kinship systems transformed. It's very simple, Terry. The serial bottleneck idea, on the other hand, assumes the progressive loss of genetic diversity in the face of the adaptive pressure to colonize the whole globe, and the corresponding accrual of linguistic and cultural diversity, plus the preservation of the ancestral types of kinship organization in the opposite corner of the world from Africa. This is total nonsense. Dienekes just wrote about it, but he, of course, has a different explanation from mine.

"Possibly. But it's greater diversity in Yemen and Oman is quite likely to be product of its arrival there from both of the regions it is found in Africa: Egypt and Ethiopia."

It's possible, just like every other reading of genetic phylogenies and diversity estimates.

onur said...

I just debunked this stereotype of massive Old World language and kinship extinctions on Razib's site.

That is a quite presumptuous statement. For this and other similarly presumptuous statements of yours, German, I have decided not to respond to your comments so often as your attitude is often self-righteous and rigid.

German Dziebel said...

"your attitude is often self-righteous and rigid."

No, Onur, people just interpret it this way because I give an informed pushback against those aspects of their thinking that they are used to freely run with without facing any criticism and without providing any proof.

German Dziebel said...

Onur,

Here's a great quote that, in a nutshell, illustrates the serious problems with the mainstream human origins story. "The case for dispersal of projectile-using humans to the Levant suffers from some of the same weaknesses as the diffusion hypothesis; namely, the lack of an artifactual “trail” linking the EUP of the Levant to another region. Fortunately, artifacts are not the only evidence for population dispersal. The hominin fossil and recent human genetic records (Grine et al. 2007; Kivisild 2007) strongly support the hypothesis that there was a dispersal of Homo sapiens populations from Africa and southern Asia to western Eurasia at around the same time as EUP assemblages began to be deposited. That the specific forms EUP projectile armatures took do not replicate African precursors does run counter to models for detecting “migration” derived from recent contexts (Clark 1994), but this is not necessarily a crucial flaw. Populations dispersing into new territories do develop novel artifact forms unknown in their donor region. For example, it is beyond serious scientific dispute that the Americas were first populated by humans dispersing there from northeastern Asia, and yet few specific artifact-types connect these two regions (Meltzer 2009)." Shea. Sisk, "Complex Projectile Technology and Homo sapiens Dispersal into Western Eurasia" // PaleoAnthropology 2010.

There's clearly no archaeological evidence for an out-of-Africa migration. Scholars acknowledge it. But, lo and behold, they use nothing else but the peopling of the New World as the justification for this lack of archaeological evidence for an out of African migration. In a perverse logic, they absolve archaeology from the need to provide material evidence for the peopling of the Americas and rely on a speculative but firm consensus that this peopling event somehow happened. This speculative consensus is then made to look like "science" and is exported back as a yardstick for the archaeological standard needed to prove the out-of-Africa model. Seeing the gap in the archaeological record, they nod in the direction of craniology and genetics. But the study of the Hofmeyer skull that they quoted actually showed that this 35,000-year-old South African specimen looks nothing like modern Khoisans, Pygmies or Negroids. It looks fully Eurasian (!) just like a bunch of Cromagnon skulls from the same period that are in fact related to modern European through straightforward morphological continuities. The genetic data they invoke, again, shows exactly the same biogeographic gap between African and Eurasian haplotypes with no "trail" connecting the two (see above). So we have a perfect match between archaeology and genetics but the one that doesn't support the out-of-Africa model but in fact directly contradicts it.

You can now see that the cross-disciplinary support for the out of Africa model is completely bogus and betrays that the whole model is patched together with rumors and not facts. This, in turn, opens up a window of opportunity for an alternative model that is based on facts such as the fact that there's no archaeological evidence for the peopling of the Americas and there's no archaeological trail for an out of Africa migration.

onur said...
This comment has been removed by the author.
onur said...

German,

As I said earlier, the archaeological record (including skeletal remains) is usually very fragmentary for so ancient times. So your conclusions aren't justified by the archaeological record. As to the Hofmeyr Skull, it seems to me that physical racial morphologies of Homo sapiens sapiens were throughout most of their Palaeolithic history in general less distinct from each other than in the Holocene and the relatively late periods of the Upper Palaeolithic. Though the sparseness and also the incomplete nature of human skull remains in many regions of the world from so ancient times can also explain this situation to some extent.

German Dziebel said...

"it seems to me that physical racial morphologies of Homo sapiens sapiens were throughout most of their Palaeolithic history were in general less distinct from each other than in the Holocene and the relatively late periods of the Upper Palaeolithic."

This is true. However, the Hofmeyer skull doesn't show any special continuities with "anatomically modern humans" from African Middle Stone Age (that should further manifest themselves in either Khoisans, Negroids or Pygmies in some form), but in fact clusters with Eurasians of the same period. So the gap is there craniologically as well.

But what I find fascinating is that archaeologists use the Hofmeyer skull - arguably the worst candidate from either your perspective or mine - as "evidence" for an out-of-Africa to patch up the gaps in their own evidence. More skulls from African Late Stone Age are of course needed.

onur said...

More skulls from African Late Stone Age are of course needed.

Yes, of course. Until then, the Hofmeyr Skull is hard to interpret.

StudiiVediceBucuresti said...

My opinion is that mongoloid race was formed in Himalaya region.It was the first cold region inhabited by humans.It explain the ancient haplo D in tibetans.It explain the bigger diversity of southern mongoloids(if the mongoloids came from central Asia or Siberia,the diversity should be bigger in the north).
It doesnt contradict the agricultural migration of mongoloids from north to Indonesia

terryt said...

"My opinion is that mongoloid race was formed in Himalaya region".

I'd place it a little north of there. Perhaps as far north as the southern sides of the mountains south of Lake Baikal.

StudiiVediceBucuresti said...

south himalaya

StudiiVediceBucuresti said...

How do you explain then,the fact that southern mongoloids have bigger genetic diversity comparative whit nordic ones?I suggested that south Himalaya was the first cold region ,near Nepal and Bhutan, that was reach by humans.

terryt said...

"How do you explain then,the fact that southern mongoloids have bigger genetic diversity comparative whit nordic ones?"

The southern ones have considerable admixture with pre-existing SE Asian populations.

"I suggested that south Himalaya was the first cold region ,near Nepal and Bhutan, that was reach by humans".

Possibly so. But people in that region are nowhere near as Mongoloid-looking as those to the north of the Himalayas.

StudiiVediceBucuresti said...

How do you know which haplos are pre-existent if any?
As far as i know,mongoloid haplos O and N have the origin in south-east Asia.Do you propose a back migration of O and N?
people from Bhutan look very mongoloid except those mixed whit indians.
Maybe they look proto-mongoloid and become more and more mongoloid as they reach new colder regions.we see that amerindians are not so mongoloid.Maybe they preserve proto-mongoloid features at the time of migration,or they lost that features in adaptation in the new american climate.
How can you know which haplos are pre-existing australoid?
What is for sure is that mongoloids spread also whit neolithic rice culture,replacing the hunter-gatherers,an least on chinese and indonesian area.

So this is the paradox of mongoloid race.Their features seem adaptation of a cold climate,yet the genetic show a more genetic diversity among southern mongoloids.An least if you can show evidence that diversity of south eats Asia is from mixing whit pre-existing genes(including haplo O and N) ,a Bhutan-Nepal origin is plausible.

Boasian said...

Beringian Isolation does not take into account Holocene back migration by Amerindians into northeast Asia. Many studies have shown that Amerindian mtDNAs require much greater time while the A,B,C, and D haplotypes need not be "founding lineages or the result of "founding Effects."

Invited Editorial, Emoke Szathmary; Am. J. Hum. Genet. 1993b pg. 796



"If Chakraborty and Weiss's (1991) findings apply in general to the Americas, it means that not only is there no evidence for the presence of major bottlenecks in the evolutionary history of the mtDNA in the New World but also that it is not possible to establish the evolutionary source of mtDNA mutations. They are as likely to be the product of new mutations as of ancient founder effects."



The only cladistic evidence of an Asian/Amerindian affinity is to be found in the discovery of the four rare Asian mtDNAs. These proposed founding Amerindian lineages (see Schurr et al. 1990; Torroni et al. 1993a), are not descendent of the derived (post-nodal) Asian lineages, those described in the trees generated by Cann et al. (1987); Johnson et al. (1983) and Excoffier and Langaney (1989). The two subgroups common to Central and Southeast Asian populations are found in 36.1% of the Siberians studied while "surprisingly" they are not found in Native Americans. Contrarily, the presence of rare Asia mtDNA in the Americas does not specifically identify that their origin must be Asian in that admixture from the Americas could be suggested for some populations inhabiting Siberia (Hicks, in submission). An alternative explanation could be proposed in the movements of Amerindians into northeast Asia following or during the formation of contemporary Circumpolar peoples in post glacial times (Boas 1905; 1910, his "Eskimo wedge theory"). Should this be the case, then any evidence for an Asian (or for that matter, African or European), origin for the Eskimos, Na-Dene, and/or Amerindian would be, cladistically, unsustainable.

terryt said...

"So this is the paradox of mongoloid race.Their features seem adaptation of a cold climate,yet the genetic show a more genetic diversity among southern mongoloids".

As I've tried to point out many times, there are many reasons why 'diversity' does not necessarily equal 'origin'.

"people from Bhutan look very mongoloid except those mixed whit indians".

Have you considered the possibility that it's the other way round? The Mongoloid people in Bhutan, Nepal and Assam may be relatively recent arrivals. In fact that is the usual explanation.

"if you can show evidence that diversity of south eats Asia is from mixing whit pre-existing genes(including haplo O and N) ,a Bhutan-Nepal origin is plausible".

But the Bhutanese-Nepalese are themselves probably a relatively recent mix of Indian and Mongoloid. Many groups of them have myths of having come through the Himalayas from the north or northeast.

"Maybe they look proto-mongoloid and become more and more mongoloid as they reach new colder regions".

To me that seems unlikely. What we see in other species is adaptation to the environment and then expansion of that phenotype.

"What is for sure is that mongoloids spread also whit neolithic rice culture,replacing the hunter-gatherers,an least on chinese and indonesian area".

Exactly. So what haplogroups might have been involved?

"Do you propose a back migration of O and N?"

How about those haplogroups as candidates?

"we see that amerindians are not so mongoloid".

Not 'so' mongoloid, but a little bit. It seems likely that they originally moved across Eurasia to the north of the Mongolian plateau, although they picked up more southerly mtDNAs along the way.

"How can you know which haplos are pre-existing australoid?"

Australia has unique haplogroups in the form of Y-hap C4 and mtDNA S. The other Australian haplogroups are shared with New Guinea, and are almost certainly more recent arrivals in Australia.

StudiiVediceBucuresti said...

Maybe diversity doesn't point the origin always ,but surely the law of probability point the origin in the region of highest diversity.Other fact s must be check.How old are the mutations,how big was population etc.
And yes ,for example central asia or turkey receive genes from all directions,but south-east asia is not in a central position,so its unlikely that it receive genes from other places. An is prety sure that it didn't receive the haplo O from India(O make the bulk of haplos there).
By ''australoids'' i didnt mean the native australians but the asian people that resemble the africans(negritos,papuans,andamans)
Now there are 2 theories of how mongoloid race developed.One say that they adapt to cold(small eyes,layer of fat,short limbs).Second theory suggest a proces of neoteeny-sation (child -like).Humans look child-like comparative whit other apes.Mongoloids look child-like comparative whit caucasoids and africans.Not only physical but also the time of learning is bigger in mongoloids(and is posible that they are smarter then Caucasoids for this reason).
These 2 theories are not mutually incompatible but if only the second is true the a stage of proto-mongoloid is possible even in a warm climate.

terryt said...

"An is prety sure that it didn't receive the haplo O from India(O make the bulk of haplos there)".

To nme O seems the only possible candidate for Neolithic move south from the Yangtze/Yellow river agricultura expansion. So it originated in that region.

"Now there are 2 theories of how mongoloid race developed.One say that they adapt to cold(small eyes,layer of fat,short limbs)".

That's the one I think is most likely, and at a relatively high altitude. Especially as they also have other adaptations to a high light-intensity environment.

StudiiVediceBucuresti said...

if you can explain why haplo O have the biggest diversity on south-east Asia and not on neolithic Yangtze river as normally should be, please share whit us.Or maybe you can prove that haplo O doesn't have the biggest diversity on south-east Asia after all.

terryt said...

"if you can explain why haplo O have the biggest diversity on south-east Asia and not on neolithic Yangtze river as normally should be, please share whit us"

Leaving aside the diversity factor for a while, we find it very difficult to postulate any possible scenario for regiona of origin, or pattern of the thre O haplogroups' movement north. On the other hand it is quite easy to postulate a pattern for their movement south. All basal O haplogroups are present round the region of the Yangtze, or nearby. O2 looks to have moved both north and south from the eastern end of the Neolithic region, O3 from the western end, and O1 from the south-central neolithic region.

Now, diversity. The Yangtze Neolithic region is relatively small, fairly long established, and numbers probably did not expand much locally for some time after the Neolithic expansion south. In such regions with stable population numbers haplogroup diversity tends to become reduced with time. But as the basal O haplogroups moved south through the mountains and valleys of South China, and population numbers through the region increased, diversity would also increase. We know that diversity is great in mountainous regions with each valley finishing up with its own haplogroup. In such regions we normally assume that diversity does not equate with origin, yet, for some unknown reason, in mountainous South China and east Asia most of us assume that diversity means 'region of origin'.

terryt said...

Further to my comment yesterday:

Perhaps it would be productive to look at the development of the Chinese Neolithic. As far as I know it began with the Yang-Shao about 7000 years ago, confined originally to the deciduous forest region of the loess-clad hill country of the middle Yellow River valley:

http://en.wikipedia.org/wiki/Yangshao_culture

Later, as it expanded onto the loess plains further east, it developed into the Lung-Shan which, according to an old Grahame Clark book I have, then expanded 'from the deciduous into the evergreen zone', and 'spread into Szechwan and south into the coastal provinces from Chekiang to Kuangsi and beyond into south-east Asia. In the coastal zone settlement was concentrated on the lower courses of rivers and on adjacent islands'.

This coastal element fits extremely well with the distribution of haplogroup O2a. Especially when we consider that O2b is also somewhat coastal, but to the north: North China, Korea and Japan (admittedly with traces down to Vietnam and Thailand).

And the Szechwan element fits a postulated centre of dispersal for haplogroup O3 very well. The haplogroup would then have spread overland down into SE Asia, and eventually out into the Pacific with the Austronesian-speaking people. It has become the most widespread and common East Asian Y-haplogroup.

Haplogroup O1 is especially common in Taiwan, the Philippines and Wallacea. But it's extremely unlikely it originated anywhere in that region. Most probably it is an immigrant to Taiwan, and spread from there at the origin of the Austronesians. To me it seems most likely that O1 originated in the middle Yangtze region, Kiangsi or Hunan, moving through Fukien and across the Formosa Strait to Taiwan.

The big population expansion in the Chinese Neolithic homeland didn't happen until the Shang state was established there. Until that time haplogropup diversity would have been progressivley reduced in the homeland.

I'm not aware of any possible scenario for the pattern of origin and expansion that fits a Southeast Asian origin for haplogroup O. Just general beliefs.