The authors of this open-access paper surveyed Kazakhs from Altai and Barguts from Inner Mongolia, comparing them with previous data of other nearby peoples and even the distant Iranians.
Miroslava Derenko et al., Complete Mitochondrial DNA Analysis of Eastern Eurasian Haplogroups Rarely Found in Populations of Northern Asia and Eastern Europe. PLoS ONE 2012. Open access.
The really novel data is of course in the list of mtDNA haplogroups found among these two populations, which is listed in table 1. While Kazakhs have 40% of Western lineages (H, U, J and others), the Barguts only display some 8% (exclusively U and HV clades).
In addition they perform a somewhat interesting comparison with other populations, mostly other Altaic-speakers:
Fig. 1 - PC analysis, color-coded for languages |
The two clusters marked in the PC1-2 graph also appear in the PC2-3 graph, what means that they are very homogeneous in fact. Maybe the Sojots were once Mongol speakers and the Turkic speakers grouped with Evenks and Todjins were once Tungusic speakers, I guess.
Also the paper provides what they claim to be a complete tree of haplogroup B4'5:
Fig. 3 - click to expand
Red: East Asian, Orange and Yellow: SE Asian, Blue: North Asian, Purple: Native American
See original for full legend |
The also provide some potentially very useful information on haplogroups R9c (including F), N9a, M9, M10, M11 and M13'46'61 (found as singletons in the sampled populations), for which they provide phylogenetic trees in the supplementary materials.
Very brief notes on these:
- R9c1 (the sibling of F) is most common in Philippines (3.3–5.7%) and Abor (11%). Frequencies decay as we move northwards.
- F2 is most common in Thailand (2.4–5.4%) and then in China (1.9–3.3%). The Bargut singleton branch is apparently a new clade (F2e proposed)
- N9a is most common in Japan (4.6%), Korea (3.9%), China (2.8%) and Mongolia (2.1%), with less important presence in Island SE Asia and Eastern Europe.
- M10 is a rare East Asian clade with very rare cases in East Europe.
- M11 is also a rare East Asian clade found from Altai to Japan.
- M13 has two branches, one (M13a) most diverse in Tibet (and often found in East Asia) while the other one (M13b) is restricted to Malaysian aboriginal peoples.
- M9 also has two branches: E is found essentially in Island SE Asia and Taiwan aborigines, while M9a'b is apparently also centered in Tibet and scattered through East Asia and Hymalayan parts of South Asia. They propose that M9a1a1 spread to North Asia from South China after the LGM (note: I do not necessarily subscribe this age estimate).
Update (May 23):
In order to illustrate the discussion (see comments) I made this map of the "Evenk cluster" which I suspect was first Evenk-speaker or otherwise ethnically homogeneous before the Turkic expansion (since less than 2000 years ago):
The most notable feature is that it seems to be more northernly or oriented to Central Siberia than the other populations. It's a Siberian-Taiga specific cluster, while most of the other Altaic populations are steppe dwellers instead. Compare with this map of the taiga:
They overlap very well, right?
Maju said,
ReplyDelete"Probably the Sojots were once Mongol speakers and the Turkic speakers grouped with Evenks and Todjins were once Tungusic speakers, I guess."
I was going to let this little error slide by, but everyone else has been nitpicking lately, and I have decided to join the bandwagon.
The Todjins are also actual Turkic speakers, Maju. The two green dots that are clustered with the Yakuts, Tuvin-Todjin-Tofalars, and Altai Kizhi represent two samples of Evenks, one "eastern" and one "western."
You are right about the Todjins. And do not worry about nit-picking, please: it's useful.
ReplyDeleteWhat do you think anyhow about my speculation on "archaic linguistics"? Does it make sense or am I threading too thin?
Sorry for moving off the subject, but good to have you back commenting Ebizur. What have you been doing?
ReplyDeleteWhat do you think anyhow about my speculation on "archaic linguistics"? Does it make sense or am I threading too thin?
ReplyDeleteIn Inner Eurasia there was never a Turkic, Mongolic, Tungusic, Iranic, Indo-European, Uralic distinction; these are just linguistic-based categories created by foreigners. When a tribe conquered and permanently settled in with some numbers a new territory it was the norm to integrate and assimilate the locals of the newly conquered territory into the conquering tribe by various means (marriages, force, alliance, slavery, etc.) irrespective of the language of the locals. So in such exogamous societies it is a futile task to search for the original Turkic, Mongolic, Tungusic, Uralic genetics based on modern populations. We can at most make good estimations about the original genetics of only some subgroups of such broad linguistic-based categories based on modern populations.
"these are just linguistic-based categories created by foreigners"...
ReplyDeleteThey are linguistic categories indeed but they are built with the lives and history of the peoples of the area. All linguists do (if they do it right) is to describe that in a methodical and scientific manner.
I feel that dismissing the matter as you do, Onur, does not honor the usefulness that this study potentially has to unravel some of the unwritten history (technically "prehistory") of these peoples.
I feel that dismissing the matter as you do, Onur, does not honor the usefulness that this study potentially has to unravel some of the unwritten history (technically "prehistory") of these peoples.
ReplyDeleteActually I am not that pessimist on this issue. Ancient DNA samples from the relevant regions and eras might unravel a lot of mysteries surrounding the genetic distant past of the peoples and regions in question. As I pointed out in my previous post, for not so distant past even the modern populations in question might give important information. The problem with inferences about their distant past based on the modern populations is that modern Mongolic, Tungusic and eastern Turkic peoples do not form distinct genetic clusters that parallel their linguistic affiliations, as there have been a lot of close contacts between them in the past. Anyway, genetics of the modern populations and spatiotemporal craniometric studies already give some coarse information about the genetics and racial attributes of the proto-Turkic, proto-Mongolic and proto-Tungusic peoples. For instance, they all were almost certainly racially Mongoloid (some possible small Caucasoid admixture in them does not change that fact).
"The problem with inferences about their distant past based on the modern populations is that modern Mongolic, Tungusic and eastern Turkic peoples do not form distinct genetic clusters that parallel their linguistic affiliations"...
ReplyDeleteActually they do in some cases to some extent, as is apparent in this study:
1. Mongol speakers (except Mongolians plus Sojots) form a tight cluster at the PC1/2 graph - and also at the PC1/3 graph but this time with Mongolians included.
2. Evenks cluster among themselves in both PCAs but they include other populations who nowadays speak Turkish: Altaians-Kizhi, Tofalars, Tuvinians, Yakuts and Todvins - suggesting that all these were once Evenk-speakers who learned Turkic upon Turkic expansion.
3. Khakassians, Altaians and Teleuts (all Turkic speakers) cluster in both graphs as well.
4. Altaian Khazaks and Telengits cluster also in both graphs, including also Sors in the PC1/3 snapshot.
5. Iranian peoples cluster even if only in the PC1/3 graph (they also do to some extent in the PC1-2 but they are much more scattered because of Altaian "oversampling" dominating both dimensions).
So we have: one Mongol cluster, two distinct Turkic cluster and one probably Evenk (Tungusic), but now largely Turkic-speaker, cluster, plus the Iranian cluster.
There is some linguistic consistence to these clusters, isn't there?
Geographically the Evenk cluster seems to be focused in the North and towards Central Siberia: it's a very Siberian cluster, be it originally Evenk-speaker in full or not.
DeleteI'll post a map as update to illustrate this point.
Maju, it is not clear where Proto-Turkic people originated. All the linguistic (whether the Altaic family exists or not), archaeological, genetic and historical evidence points to somewhere in the northeast of Asia, But it is not clear in which part of NE Asia they originated. The earliest Turkic peoples might as well be hunter-gatherers as the earliest Mongolic and Tungusic peoples seem to be. Pastoralism might be a secondary development for Turkic peoples, perhaps as a result of the contacts with the easternmost Iranic pastoralists of Inner Eurasia. Siberian taiga zone is not excluded from the regions where proto-Turkic people might have originated.
ReplyDeleteI don't want to reach to any conclusion: all I say is that this particular cluster #2 includes non-Turkic-speakers and it is less likely that Evenk language expanded on a previously Turkic cluster.
DeleteThe two Turkic-only clusters are one (#3) from further south (Central/South Altai, Kazakhstan) and the other (#4) from a very specific area of the Taiga NE of Altai (Khakassia & Kemerovo Oblast).
And then you have the Mongol cluster (#1) in, essentially, Mongolia-Buryatia.
However I was told years ago (by Turks interested in history) that it is likely that the original Turkic core population was from modern-day Mongolia (much of it, not all) and core Mongols from further East, possibly the Shiwei, at least in part.
The Altai region, the western third of Mongolia, southwestern Siberia and of course Central Asia (including Xinjiang) are unlikely to be the Turkic urheimat, as these regions were all populated by Caucasoid Indo-European-speaking peoples until the Turkic expansion (which began with the Xiongnu and intensified with the Turkic Khanate). The middle and eastern thirds of Mongolia (also Inner Mongolia and Buryatia) are of course among the candidates for the Turkic urheimat. But Manchuria and the eastern half of Siberia (except its far north and northeast parts) too.
DeleteThis comment has been removed by the author.
ReplyDelete