Thursday, March 22, 2018

The whiteboard


David Reich's book, Who We Are and How We Got Here: Ancient DNA and the New Science of the Human Past, is coming out next Tuesday (see here). Chapter 6 has the potentially controversial title The Collision that Formed India, and indeed I know for a fact that Bronze Age steppe pastoralists, who seem to induce panic attacks amongst a lot of people, and especially Out-of-India proponents, get a big hat tip in this chapter.

But I can't really say more than that until after the book launch. So in the meantime, let's focus on this intriguing photo of a messy whiteboard that was published in the New York Times this week along with a feature on the Reich Lab's work with ancient DNA. The version below was edited by me to highlight and fill in a few details. The original can be viewed by scrolling down here.


Clearly, this is a mixture and migration model for Central Asia and surrounds covering the crucial period 2200-1500BCE, when, according to a consensus amongst historical linguists, waves of Indo-European speakers moved into the region from the steppes. It's probably from a jam session about an upcoming ancient DNA paper. Here's my interpretation of the model:

- nodes 1, 2, 3 and 4 track the migration of Bronze Age pastoralists from the Pontic-Caspian steppe deep into Central Asia, while nodes B and C follow the expansion of Neolithic farmers from east of Anatolia (probably from somewhere in present-day Iran) into Central and South Asia (nodes 1 and B aren't actually visible in the original pic, but must be there, and more or less where I marked them)

- node 2 probably represents the formation of late Corded Ware Culture (CWC) populations across Northern Europe around 2900 BCE, via the mixture of Yamnaya or Yamnaya-related steppe pastoralists (node 1) with European farmers, who were themselves a mixture of Anatolian farmers and Western European Hunter-Gatherers (WHG)

- Sintashta and Andronovo_NW at node 3 derive directly from the mixture event at node 2, so either they're offshoots of late CWC or a closely related population

- intriguingly, and perhaps crucially, nodes 2 and 3 only take one pulse of admixture from node 1 (red X), while the branch leading to Andronovo_SE at node 4 takes two such pulses, with one apparently later than 1900 BCE, possibly suggesting that Andronovo_SE was more Yamnaya-like compared to late CWC, Sintashta and Andronovo_NW

- moreover, the branch leading to Andronovo_SE absorbs significant admixture from Western Siberian Hunter-Gatherers (West_Siberian_HG) and possibly a Central Asian ghost population, no doubt resulting in a further reduction of Anatolian farmer and WHG ancestry ratios in Andronovo_SE compared to Sintashta and Andronovo_NW

- thus, Andronovo_SE, unlike Sintashta, might fit the bill statistically as enough Yamnaya-like to be the Yamnaya-related steppe pastoralists who "crashed" into India during the Bronze Age (see here), although, admittedly, this isn't actually shown on the whiteboard

- on the other hand, if, perhaps, the model includes a migration edge from node 1 to B, then this would suggest that Yamnaya-related ancestry arrived in South Asia with a very different population than Andronovo_SE, and possibly much earlier than 1500 BCE, but we don't know because David Reich is (strategically?) blocking that part of the whiteboard.

Also worth noting is that there's actually nothing about India in the model. The most proximate region that gets a mention is "Turan/Northern South Asia". So should we be concerned that the supposedly imminent publication of ancient DNA from Rakhigarhi and other Indian prehistoric sites has been pushed back indefinitely, perhaps for political reasons? Normally I'd say no, but in recent weeks I've been hearing rumors that this is indeed the case.

Update 01/04/2018: The preprint of the paper on ancient Central Asia that I mentioned above is now available at bioRxiv. See here.

See also...

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

114 comments:

  1. He has a sly grin and is deliberately standing in front of the key part of the chart. I think he is having a little joke with us.

    ReplyDelete
  2. Finally some news.

    Am I the only one who thinks that the green arrows below Andronovo and above Turan represent BMAC?

    PS. I'm glad my theory of Indo-Iranians coming from Armenia is being pushed here :P

    ReplyDelete
  3. It would be interesting to assign languages to each of the nodes. I assume each of the nodes 1,2,3 and 4 was IE. But if there was a a migration from node 1 to node B, then B may be IE too. In such a case the history of IE languages may be much more complicated than we used to think. Let’s wait for the paper before we start speculating

    ReplyDelete
  4. Before this study comes out (which it soon, likely will), I wish we had EXTENSIVE sampling of different Yamna archaeological groups with their associated genetics. Treating the Yamna as a single group isn't solving all the problems. There is a hell of a lot of clades and subclades within Yamnaya itself to skip this.

    I wish someone from one of the genetics teams would read this or other comments expressing the same desire.

    ReplyDelete
  5. The title on Indian chapter is bold :) , but it speaks on ANI and ASI admixture which formed modern S Asians most likely .

    The Turan in that picture is BMAC and perhaps more Northern regions .

    I don't see arrows going to India from Steppe BTW ...ding ding ding...

    ReplyDelete
  6. There was an arrow going from Sintashta (node 3, 1900BCE) into 'Turan', aswell as something coming from the East. Both got erased into the current model on the board.
    The cryptic green arrows seem to radiate all the way to 3->4 and B->C, which might indicate several continuous admixture events over the centuries, but we don't really know what they represent (unless it's all W_Siberia HG, although I find it hard to believe the existed inbetween Andronovo and the Southern route guys A->B->C.

    Nirjhar, there's no obvious India on the board, unless it's C/Turan. Either way, Reich is obviously not totally showing his hand, that would probably not be good for his book sales, best to give a sneek peak and leave the good stuff for those who pay for it. That's what I'd do anyway

    ReplyDelete
  7. @Nirjhar Because of the climate mainly. And because archaeogenetic studies in India are yet to pick up pace. But the second issue may be solved in the future

    ReplyDelete
  8. I feel the second pulse David is talking about is a Yamna/Poltavka like population which got assimilated into the Sintashta->Andronovo wave as it swept through Central Asia.

    ReplyDelete
  9. One more prediction. If there is a node from 1 to B, then that would mean some kind of steppe admixture arrived in India perhaps around 2000 BC. But that still wouldn't mean the Rig Vedic Indo-Aryans arrived at this time since frankly, there isn't any evidence in the Rigveda for urban living that we would normally attribute to the IVC

    ReplyDelete
  10. "So he's saying Sintashta and Andronovo have nothing to do with the spread of IE languages into Iran and India."

    India doesn't figure in the picture. But obviously what it seems to hint at is that South-East Andronovo did have a role in spreading IE languages into India. But that south east Andronovo was somehow more Yamnaya like than the northern Sintashta Andronovo.

    ReplyDelete
  11. Ancient Fennoscandian genomes reveal origin and spread of Siberian ancestry in Europe

    https://www.biorxiv.org/content/early/2018/03/22/285437

    ReplyDelete
  12. Nirjhar, I think the red arrows are showing a contact zone between the Turan Plain and the North Indus. Behind his back is probably Anatolian/CHG which appears to have already spread into this contact zone as another independent migration. (?)

    ReplyDelete
  13. Turan is sometimes used as name for Iranian Central Asia, centered around and beyond the Oxus. BMAC is centered around the Oxus AFIAK. Torun/Northern South-Asia is BMAC and Swat.

    ReplyDelete
  14. Vara , give your opinion on Turan please....

    ReplyDelete
  15. Reich has 1500 bc swat and 2000 bc Uzbek samples presumably pre fedorovo. He does not seem to make a connection between these and andronovo

    ReplyDelete
  16. @Nirjhar I thought C is Turan/ North South Asia. Probably referring to the BMAC/Oxus culture.

    @Jijnasu One does not need to go as extreme as Parpola to accept that there might have been some kind of steppe presence in Central Asia before and around 2000 BC, but the Vedic Aryans came later. But of course, Parpola's idea of a separate "Atharvavedic" ethnic group is without evidence.

    ReplyDelete
  17. @Chetan, you wrote:"Parpola's idea of a separate *Atharvavedic* ethnic group is without evidence".

    Actually Parpola refers to an archaeological discovery in Indus valley of axes apparently from 2000 BC to support his atharvavedic hypothesis.

    ReplyDelete
  18. @Chetan

    " there isn't any evidence in the Rigveda for urban living"

    I know the Rigveda is too long and nobody likes to read it but what you said is simply not true.

    Cereal Farming:

    "And may he. duly bring to me the six bound closely, through these drops,
    As one who ploughs with steers brings corn." Mandala 1, H23

    "7. The groats have we prepared for thee with Pusan, corn for thee, Lord of Bay Steeds, with thy horses.
    Eat thou the meal-cake, banded with the Maruts, wise Hero, Vrtra-slayer, drink the Soma.
    8. Bring forth the roasted corn to meet him quickly, cake for the bravest Hero mid the heroes.
    Indra, may hymns accordant with thee daily strengthen thee, Bold One, for the draught of Soma." Mandala 3, H52

    There are mentions of houses in Mandala 2, the oldest book according to Witzel, and battles between Devas and Asuras in 100 walled fortresses/cities. I do not think the Indo-Aryans were worshiping the natives.

    ReplyDelete

  19. @Chetan,

    You can find Parpola`s article at:

    https://tinyurl.com/y8n9nklf

    ReplyDelete
  20. @Carlos First of all, I'm not sure which paper of Parpola's you are referring to. Is it in the one which came out recently, about Indo-Iranian -Uralic contacts? I'm yet to read that.

    Also, the general view of Atharvaveda within India is that it is a "folk" Veda dealing with charms and magic in contrast to the other 3 "high" Brahmanical Vedas. That is the reason it is often separated from the other three. But if Parpola has a new hypothesis about, I must surely give it a read.

    @Nirjhar I think the linguistic arguments from him have strength and also consider that the Indo-Aryan identity, can't be fully represented by Vedic alone?.

    Yes for sure. I tend to think of the Indo-Aryan migration as consisting of 2 elements - a folk element which is the migration of the masses and an orthodox Vedic cultural element. The latter was mainly concentrated in the Kuru-Panchala region after about 1000 BC. So like you said outside of this orthodox region, there must have been other dialects and cultures of Indo-Aryan mixed with native traditions.


    ReplyDelete
  21. @Chetan

    I posted the link above. I will do it again:

    https://tinyurl.com/y8n9nklf

    Parpola, Asko, 2015. The Mohenjo-Daro axe-adze: A vestige of Aryan immigrations to Central and South Asia? Current World Archaeology issue 74 = vol 7 (2): 14-15.

    ReplyDelete
  22. @Nirjhar007

    Turan stands for BMAC and Northern South-Asian for Swat, is what I meant. "Torun" was my spelling mistake.

    ReplyDelete
  23. @Vara I never claimed that the RigVedic people were unfamiliar with agriculture. I said they don't seem to have been an urban civilization like IVC.


    @Carlos Parpola says Tepe Hissar was occupied by people who were familiar with the chariot from 2000 BCE. And their axes have also been found in the top layer of IV. By means of trade or small scale migration probably

    ReplyDelete
  24. @Carlos Actually that situation is looking very probable now. Some kind of Yamna related migrations into central Asia which happened before the Andronovo expansion. Probably this Tepe Hissar culture got assimilated into the south-east Andronovo like the diagram shows.

    ReplyDelete
  25. @Nirjhar Actually, I got the impression that India is not shown in the diagram at all. Turan is probably referring to the Afghan region.

    ReplyDelete
  26. I think the red arrows are ASI and node c combines uzbek and swat. Node 1 takes on some Iran-neo before impacting the northern nodes as in lazardis

    ReplyDelete
  27. I don’t even bother reading parpola. Seems like completely blind conjectures

    ReplyDelete
  28. @Chetan

    The article mentions Uzbek samples. I reckon they are referred to as "Turan", as that area was tied to the Oxus river, the current day Amu Darja.

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

    ReplyDelete
  29. @epoch Yes they're probably BMAC samples

    ReplyDelete
  30. @Nirjhar

    A promise is a promise :)

    1: Historical evidence

    The majority view is that Turan is northeast of Iran comes from the late Sassanid Period, in which their Hephthalite and Gokturk enemies were considered the Turanian descendants of Franrasyan. However, Ardashir I, the founder of the empire, definitely disagrees with that as he puts Turan between Makran and Kushan. His grandson, Narseh, was the governer of Sindh, Sakastan and Turan, this area is later considered part of the Nemroz(South East) region. The early Sassanids are considered to be more reliable in terms of their knowledge of the Iranian history, indeed for atleast they knew of the Achaemenid Empire and Achaemenid names which became extinct in Iran after the 6th Century, while the Late Sassanid Kings changed history and myths to fit with their own image when they wrote the Khwaday Namag.

    2: Myths

    These aren't reliable as you know. The location of the Iranian homeland changes from one text to the other depending on the period in which the text was composed. However, as I've told you before the most likely homeland is around the Helmand.

    Turanians were mentioned only once in the Gathas where one of them helps Zoroaster. In later texts there are made to be the main enemies of Zoroaster while in the Gathas his enemies were other Priest Aryans. Sadly, they were only mentioned in the later Yashts with their location being unclear.

    However, there is one Turanian figure/family whose location is consistent in all the myths; Keresaspa/ the House of Saam. During Ferdowsi's time Rustam and his father Dastan, who were originally Saka heroes, became the heirs of Kersespa and heroes of Iran, while in Pahlavi texts and even in Garshaspnama they were of Turanian descent. I think even some parts of the Shahnameh alludes to the fact that Zabolestan was a separate country from Iran, especially in Rustam and Isfandiyar and Rustam and Sohrab. Anyways, in all versions of the myth Kersespa, Saam, or Rustam were anti-Zoroastrian kings or governers of Nemroz (mostly Zabolistan) making the early Sassanid version correct.

    ReplyDelete
  31. I actually think B is an Iran_Chl or Iran_N like population, and C is the BMAC, which received major ancestry from Iran_Chl (black arrow from B to C) and also ancestry from the Indus Valley (red arrow from South Asia) and also underwent genetic exchange with Andronovo (green arrows) leading to an admixture bar for 4 (late Andronovo) with a minor fraction of ancestry by 1300 BC. That makes most sense to me. But we'll see.

    He's also consciously trolling people. Certainly he's succeeded, must be having a good laugh with his lab-mates when he saw his silhouette...

    ReplyDelete
  32. South Asia generally refers to the Indian subcontinent. So Northern South Asia referred to in C should be North India right? Doesn't this board clearly show that Sintashta, Andronovo are a separate branch from Turan/North India? Wouldn't this debunk the whole chariot riders from Sintashta/Andronovo crashing into North India after spending sometime in BMAC which is the current consensus of the Aryan Migration theory?

    ReplyDelete
  33. @ryu
    The green arrows are west Siberian hg common to both node c and se andronovo and red is something different from that. Goin by article they sampled Uzbekistan. can that be called bmac?

    ReplyDelete
  34. David Reich comes off to me as someone not afraid of controversy. He's a brainiac who wants to find the truth no matter what and is driven by a genuine passion for human genetic history.

    Recall, he didn't compromise for German researchers who thought his conclusions on Corded Ware were somehow in line with Nazi propaganda.

    I can't imagine he would postpone a paper on south Asian ancient DNA or shy away from the Indo European question because of controversy.

    ReplyDelete
  35. That's why I'm excited for David Reich's new book. He won't twist the data to fit a conservative pro-migrationist, racialists or liberal anti-migrationist, race-denialist narrative. He'll report on whatever the data shows.

    ReplyDelete
  36. @postneo

    "Goin by article they sampled Uzbekistan. can that be called bmac?"

    BMAC is from Bactrian-Margiana Archaelogical Complex.

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

    ReplyDelete
  37. Yes I know ..my main point is that and most of eastern activity on the white board is late or post bmac. Parts of Uzbekistan did overlap with bmac but it’s not a given that the samples are bmac proper.

    ReplyDelete
  38. @ Samuel

    Scientific trolling is best trolling, definitely. A little bit of willingness to be disagreeable seems useful, and humour is always good :)

    ReplyDelete
  39. @postneo

    We're talking on the basis of the picture in a newspaper of a man in front of a white board, so I thought I could permit myself some broad sweeping speculations ;)

    ReplyDelete
  40. If you look at the way his lab has done things, they always use aDNA to triangulate the existence of a ghost, then produce a next paper that fills up that ghost with a real aDNA sample. Its a case of stepping stones, first "three populations founded Europe" with a Native-American like ghost, which was filled with ANE-Mal'ta, then 'Q' (or the much less poetic 'Teal', what a terrible name to be cursed with) which was CHG and Iran_N.

    In this case they may have reversed this order, with the Southern Andronovo+"Turan" samples triangulating with Iran_N and Onge for India.

    Maybe they will also release Maykop and Kura-Araxes with this bunch, especially if these show influences from the BMAC.

    @ Postneo

    He's most definitely indicating exchange and infiltration between Andronovo and "Turan".

    ReplyDelete
  41. Whichever date and direction the migration into South Asia was, it is clear from the literary texts that Saraswati was a mighty river during the early Mandalas of the Rig Veda, and later texts describe the river disappearing and reappearing at multiple locations and also drying up in the desert (Mahabharata). Even later texts no longer concentrate on Saraswati, but on Ganges.

    Imperial College London study shows that Sutlej last flowed into Ghaggar Hakra paleochannel 8000kya, before changing course. Monsoon started declining 7000kya (IIT Kanpur study). You see maximum number of excavated Indus-Saraswati settlements along the Ghaggar paleochannel, some of them the biggest.

    IMO, theres strong evidence that Rig Vedic people are much older than thought.





    ReplyDelete
  42. Also does it say Anatolia under A and A seems to be the only node without an in-arrow? Does this mean Reich believes the PIE is Anatolia? I'm confused.

    ReplyDelete
  43. @Ryu,
    "Scientific trolling is best trolling, definitely. A little bit of willingness to be disagreeable seems useful, and humour is always good :)"

    Ummm...But I don't think Reich is trolling.

    ReplyDelete
  44. The nMonte Global25 analysis for Uralic, Turkic and Iranic populations of Central and North Asia is done. Link here.

    ReplyDelete
  45. if someone can help - On what evidence has the floor age of the PIE people been established? what is the floor?

    ReplyDelete
  46. Is that third population(?) in the small graphic named Tyumen i.e. western Siberia?

    @David,

    Since the topic of re-sorting populations came up in the previous thread again, do you think you could move NA17374 and S_GREEK2 to the Trabzon ref (or all together to an East Anatolian one) and NA17376 (I wonder about NA17375 too but it's not as absolutely clear cut), maybe along with Crete, to an Aegean one?

    ReplyDelete
  47. The exchange between “Turan” (BMAC) and andronovo has long been expected
    But what about further on ?


    And what’s that additional migration edge from (1) to (3 1/2) ?

    ReplyDelete
  48. This comment has been removed by the author.

    ReplyDelete
  49. @ David

    Possible to give us a bit more info abt the "pushed back indefinitely" thing?

    ReplyDelete
  50. @Salden,

    >So tell us how European admixture or close to it ended up in India (see R1a) but not no South Asian admixture ended up in Europe or areas near it.

    Obviously there's the possiblity of South Asian admixture happening long after any migration type event. I wonder if node 1 is a population containing Z282 and node B containing Z93? That would mean the Indo-Iranians (node B) and the Europeans (node 1) split off before 2900 BC (according to Underhill it was 5800 years ago). Reich has shown that South Asian admixture happened between 2000 to 4000 yrs ago which would put the admixture event in 2000 BC at the earliest.

    ReplyDelete
  51. How so ryu?

    Node c is a late or post bmac population and it primarily derives from b. Supposedly this is when “Turan” should have been overrun by andronovo chariots. We see some independently shared hg like component between them but nothing more significant. Otherwise it would have been drawn explicitly

    ReplyDelete
  52. I still not totally sure about this "Turan" label... if the Andronovo south basically reaches into the southern reaches of the described Andronovo horizon, then that area basically is "Turan"* and the Swat Valley Culture is not really too far away from that as it is, which leaves little space for a Turan "zone". But perhaps this is so. adna scientists enjoy redefining geography in any case ;).

    I do think the lines pointing at the paths leading to Andronovo_S and C probably indicate mutual influence between them rather than a HG substrate like Keltiminar or something... I'd be pretty happy with more genetic "persistence" of ancient people from the Central Asian region. Frankly it would be a more interesting story, and the survival of EHG/ANE like people as the main denizens of Central Asia east of the Arals, and more than that as actual contributors to present day people, and a little understood region of the prehistoric world being genetically important to the present, would all be cool. But there's no labeling that indicates that to me. (West Siberian HGs are labeled on there... some Central Asian HG is not really).

    *That is, see - https://upload.wikimedia.org/wikipedia/commons/thumb/b/b8/Indo-Iranian_origins.png/300px-Indo-Iranian_origins.png for the limits of the Andronovo horizon and https://en.wikipedia.org/wiki/Turan#/media/File:Iran_Turan_map_1843.jpg for "Turan". Not much of a Turan zone that doesn't overlap Andronovo_S, and if "Turan" is simply BMAC why not label it as so?

    But this is slightly unnecessary Kremlinology of course. (And to be frank it's a graph thrown up in quick time rather than a perfect work of representation... there's not even a line between the Middle Neolithic and Anatolia Neolithic!).

    ReplyDelete
  53. @ Alogo
    LMAO
    Dude, if it does turn out we have genomes from Tuymen in Russia, this is a triumph of autistic internet sleuthing of the highest order man. We are so damn obsessed its funny sometimes.

    ReplyDelete
  54. @Chetan,

    You wrote: "Actually that situation is looking very probable now. Some kind of Yamna related migrations into central Asia which happened before the Andronovo expansion. Probably this Tepe Hissar culture got assimilated into the south-east Andronovo like the diagram shows".

    I agree with a possible early migration to Southern Central Asia reaching South Asia. And I think of a possible Indus Civilization`s genetical component from the Steppe maybe since c.2500 BC within multicultural Harappans (in the way J.M.Kenoyer already proposed). My contribution, by now, is to see Harappans` likely attitude towards Steppe or Southern Central Asians with Steppe origins in the same way as Romans regarded Germanic people as "barbarians" even though both shared the same Indo-European milieu.

    ReplyDelete
  55. @zulla
    In all probablity the the Ghaggar-Hakra was a slow flowing rain fed river system) in the 3rd millenium. Increasing aridity in the succeeding millenium probably led to most of it drying up. ( The exact changes in rainfall patterns are controversial and some studies show contradictory results.). There is no need to invoke a glacier fed ghaggar-hakra 7000 ybp to explain the rig vedic saraswati.

    ReplyDelete
  56. We have information that Reich thinks that the Aryans invaded around 2500 BC. not 1500 BC. See the following tweet from a former member of his lab who must be privy to the thinking of that group. He also speculates on a connection between the diversification of the Dravidian languages which has been dated to 2500 BC. by a recent Max Planck Institute study and the invasion of the “Aryans”.

    https://twitter.com/vagheesh/status/976279589243707394

    This means, of course, that they believe that the Andronovo (NW or SE) were not the invading Aryans. As @E.D said it would “debunk the whole chariot riders from Sintashta/Andronovo crashing into North India after spending sometime in BMAC”. It will also mean that the Harappan civilization was Indo-European and was created by the invading Aryans!

    We will know for sure next week when the book is released and the paper is published.

    In the mean time, I would like to draw attention to a paper which casts doubts on AIT. Palanichamy et al. found U5a1a1 in low and middle caste Dravidian speakers in south India.

    https://link.springer.com/article/10.1007%2Fs00439-015-1547-4

    But U5a1a1 is a fairly young mtDNA clade which even if it arrived with invading Aryans would not be expected in Dravidian low or middle caste people.

    ReplyDelete
    Replies
    1. Can we stop calling it an invasion until we have some hard archaeological evidence? If the date is indeed pushed back to 2500BC when the Indus valley was in full swing and war chariots weren't invented yet invasion seems extremely unlikely.

      Delete
  57. @ ryukendo
    Try Baltic_BA in your models. Distances will improve.

    Unscaled. Before:
    Finnish_East
    Nordic_IA 54.35%
    Comb_Ceramic 21.3%
    Scythian_AldyBel 12.85%
    Nganassan 7.25%
    Sarmatian_Pokrovka 3.95%
    Srubnaya_outlier 0.3%
    Distance 2.4603%

    After:
    Finnish_East
    Baltic_BA 39.2%
    Nordic_IA 33.45%
    Mezhovskaya 19.25%
    Nganassan 5.9%
    Srubnaya_outlier 2.2%
    Distance 1.3759%

    ReplyDelete
  58. @Salden

    Reminder: There isn't a single comment here proposing a migration from India to Europe, it's B --> 1.

    ReplyDelete
  59. The latest carbons dates cast doubt on andronovo expanding simply from north(west) to south (east)

    ReplyDelete
  60. We can be very confident the Bolshoy Oleni Ostrov genomes with Y DNA N1c were early Finno Urgics. Or at least representative of the population who spread Uralic languages into Fennoscandia.

    https://www.biorxiv.org/content/biorxiv/early/2018/03/22/285437.full.pdf

    In the Eurasian PCA on Figure 2, modern Saami cluster perfectly inbetween northern Europeans and Bolshoy Oleni Ostrov. One of the Iron age Finn genomes apparently has no Asian ancestry and clusters with modern Lithuanians (JK2065). Saami look more or less like a 50/50 mixture between JK2065/Bolshoy Oleni Ostrov.

    Bolshoy Oleni Ostrov had no or very little Steppe or EEF ancestry. They were basically a EHG/Siberian mix. Look what happens when using Global25 Saami are as a mixture between northern Europe, SIberia, and Mesolithic-Neolithic genomes.

    2.8274"

    Saami

    Lithuanian,31.3
    Norwegian,22.1
    Nganassan,18.6
    EHG,17.5
    Yukagir_Tundra,6
    SHG,4.3
    Polish,0.2
    AfontovaGora3,0
    MA1,0
    CHG,0
    Iran_N,0
    WHG,0

    So, yeah, this model makes a lot of sense. Overall, this means Saami may be as much as 40% EHG! That's a lot. Yamnaya was about 60% EHG.

    Results for Finns. Notice EHG & Siberian scores are the same, which is consistent with early Finno Urgics being an even EHG/Siberian mix.

    1.4684"

    Finnish

    Lithuanian,49.4
    Norwegian,38.1
    EHG,6.8
    Nganassan,5.4
    SHG,0.3
    AfontovaGora3,0
    MA1,0
    CHG,0
    Iran_N,0
    Yukagir_Tundra,0
    WHG,0
    Polish,0

    1.4872"

    Karelian

    Lithuanian,51.1
    Polish,15.8
    Norwegian,13.8
    EHG,8.8
    Nganassan,7.1
    SHG,2.7
    Yukagir_Tundra,0.7

    ReplyDelete
  61. Samuel Andrews said...
    In the Eurasian PCA on Figure 2, modern Saami cluster perfectly inbetween northern Europeans and Bolshoy Oleni Ostrov. One of the Iron age Finn genomes apparently has no Asian ancestry and clusters with modern Lithuanians (JK2065). Saami look more or less like a 50/50 mixture between JK2065/Bolshoy Oleni Ostrov.
    Bolshoy Oleni Ostrov had no or very little Steppe or EEF ancestry. They were basically a EHG/Siberian mix


    It was so predictable.
    https://s6.postimg.org/s6q0eztmp/Uralic-_Baltic-_Slavic.png

    ReplyDelete
  62. @ Sam

    The early Uralics were nothing like Bolshoi people, in fact they were far more similar to Sintashta and Andronovo-type peoples in the Fatyanovo, Balanovo, Fedorovo and Mezhovskaya cultures. You have to update yourself on the ethnographic history and linguistic evidence regarding the Uralics, they certainly did not expand from East to West, nor were they Nganassan-like at the proto-language level when they were still one community.

    We have aDNA from Mezhovskaya, which is basically Andronovo + a hefty helping of local Siberian ancestry, and in nMonte they share a very large quantity of drift with Uralics and contributes to them consistenly. In fact they are virtually a stand-in for Saami, and probably were quite close to Iron Age Saami by extension.

    [1] "distance%=1.8834 / distance=0.018834"

    Mezhovskaya

    Srubnaya 60.40
    Srubnaya_outlier 17.65
    Narva_Lithuania 7.40
    Nganassan 7.35
    Karasuk_outlier 7.20

    Nganassan themselves were only Uralicised in the 2nd mil CE, prior to this they had connections to Turkic, Evenk, and especially Yukagir peoples. the Mansi-Khanty area and especially Samoyed area, with the most Nganasan and least European-like genotypes, were also some of the later regions to be Uralicised.

    ReplyDelete
  63. @Ryu,
    "We have aDNA from Mezhovskaya, which is basically Andronovo + a hefty helping of local Siberian ancestry,"

    7% Nganassan is not much Siberian ancestry.

    "In fact they are virtually a stand-in for Saami, and probably were quite close to Iron Age Saami by extension."

    What do you mean by stand in? They aren't similar to Saami. They have 3x less Siberian ancestry, at least x2 more Steppe ancestry, less EEF ancestry.

    ReplyDelete
  64. @Ryu,

    Modern Uralics carry Bolshoi Y DNA in the same way modern Indo Europeans carry Steppe Y DNA. Saami are intermediate between Bolshoi and northern Europeans. Like how northern Europeans are intermediate between Steppe and EEF/WHG.

    To me, it's pretty obvious what's going on.

    ReplyDelete
  65. @Ryu,
    "Nganassan themselves were only Uralicised in the 2nd mil CE"

    Do you mean the region Nganassans currently live in was settled by Uralic speakers, their ancestors, recently? Because it is impossible for a people group to be Uralicised.

    Nganassans are by definition a Uralic people. They were never Uralicised in the same sense Russians were never Slavized. There is no Russian without Slavs, no Nganassans without Uralics.

    ReplyDelete
  66. @Ryu,

    A scienero with a Steppe/Siberian mix coming into Fennoscandia would demand there being an EEF_rich population living in Fennoscandia which looks unlikely. Even Saami have significant EEF ancestry, possibly at 20% Anatolian farmer. They get this from their standard IE European ancestry coming mostly from the south Baltic.

    Fennoscandia Uralic(s) are normal northern Europeans with significant Bolshoi-like ancestry. No signs of excess Steppe.

    ReplyDelete
  67. @ Sam

    Nganasans are hunter-gatherers, its obvious that a people like them could never be the progenitors of Uralic speakers, who were familiar with metal at the proto-language level and who, excepting Nganasan, all have modern European ancestry (i.e. with EEF and CHG) and who all practice farming or pastoralism of some kind, even if reindeer herding only. Nganasan do not practice even that, and the stories we have recorded from them in the 17th century makes it quite clear they were originally Hunter-gatherer Evenks and Yukagirs who assimilated into the Samoyed identity just a few centuries before.

    When modelling Uralics, whenever Saami are pushed out, Mezhovskaya+Nganasan appear in their place at a 3:1 ratio. This ancestry combination peaks in the Volga Uralics, and mixes with Nordic_IA and Combed Ware in Finnics and Saami and with Scythian, Nganasan and Yukagir ancestry as we move East.

    The expansion of the Uralic languages is associated with star-shaped patterns in Y Hgs and movements of Steppic ancestry onto substrates of Siberian ancestry, not unlike the Turkic expansions into Siberia, which created peoples like Yakuts and Dolgan, who resemble Uralics; like them, these groups abandoned horses and cattle and started focusing on reindeer. It is not consistent with a hunter-gatherer phenomenon.

    ReplyDelete
  68. Saami are also hunter gatherers. They're like white eskimos. Their lifestyle & religion & even their singing resembles Siberians. Like I said, pretty dang obvious what's going on.

    ReplyDelete
  69. @Sam

    They're like white eskimos.

    This is totally untrue.

    Just because most Uralics today live relatively simply doesn't mean this was always the case. Just look at the Yakuts.

    If you think the proto Uralics were Nganasan like Hunter gatherers from the Tundra who somehow moved west to assimilate metal-age European agropastoralists, you've got it completely backwards.

    ReplyDelete
  70. @Balaji "We have information that Reich thinks that the Aryans invaded around 2500 BC. not 1500 BC"

    I think 2500 BC is not even on the timeline. The best case scenario is an early movement (2200-2100 BC) movement from 1 (Yamna/Poltavka) to B (Oxus) giving rise to a Indo-Aryan speaking BMAC culture in the northern reaches of Afghanistan.

    But frankly I will be very surprised if these were the Rigvedic Aryans who moved into India around 1500 BC. Because there is nothing in the Rigveda that indicates its composers were coming from a long urban stay outside India. In fact Rigvedic Aryans prided themselves on being the breakers and destroyers of forts and urban dwellings.

    Maybe the Rigvedic Aryans came as a second wave with Andronovo_SE and the BMAC were the famed Dasas?

    ReplyDelete
  71. Also if those dots on the timeline are supposed to mean anything at all, we can be sure of one thing. There are no Swat samples in this paper. Because the last dot is at around 1600 BC.

    ReplyDelete
  72. I see one last possibility. That the movement 1 to B was pretty early- in the early stages of Yamna (3300-3000 BC), but in that case, we shouldn't expect the language to have been Proto Indo-Aryan or even Indo-Iranian at all. It would have been an archaic language similar to Tocharian.

    ReplyDelete
  73. @Ryu,

    Uralic(s) in Russia look so Andronovo-like for the same reason Estonians look so Baltic BA like. It isn't because early Uralics had a lot of Andronovo ancestry.

    ReplyDelete
  74. Would be interesting to view the document in the URL at the top left corner. I make it out as "data1/hms/genetics/reich/1000Genomes/om?_supp..."

    I looked through the public Reich and Broad repositories but nothing like that directory structure shows up.. it's probably an intranet, or LAN path anyway.

    ReplyDelete
  75. @jijnasu

    There are settlements in Banawali & Kalibangan built right inside the kilometres wide Ghaggar paleochannel. Banawali settlements have been dated 2500-1700Bc. It is evident that no mighty river flowed there at that time, and the monsoonal river seemingly posed little threat to the settlements.

    All scholars agree that RigVeda describes the Punjab region.Given that Banawali is in the heart of IVC settlements in Haryana, and not downstream like Kalibangan in Rajasthan, there is solid evidence that Rig vedic people lived earlier than thought.


    ReplyDelete
  76. @ Sam

    You obviously know nothing about Uralic ethnogenesis; the least you could do is to look it up, and stop spreading misinformation in the meantime.

    The obvious change between Bolshoi people and Iron Age Saami is an introduction of ancestry similar to modern Europeans, with CHG and EEF. If Mansi-like populations extended from Central Siberia to Fennoscandia, then the change from then to now is a Europeanisation of the gene pool throughout European Siberia, including among Uralics.

    Just because Nganasan-like ancestry was spread by proto-Uralics does not mean proto-Uralics were purely Nganasan-like. Thats like saying proto-IEs were an EHG-only population. CHG and EEF were important in IE ethnogenesis (in fact most of the material culture was contributed by CHG and EEF-like peoples). Similarly, the autosomes of Steppic peoples were very important in proto-Uralic ethnogenesis, in keeping with the linguists' placement of the Uralic homeland close to early IE populations in the Volga and surrounds. Your interpretation is just incorrect.

    ReplyDelete
  77. @chetan why couldn't the language be proto Indo aryan or Indo Iranian by that time?

    ReplyDelete
  78. This comment has been removed by the author.

    ReplyDelete
  79. @Mr Kulkarni 3500-3000 BCE is considered within the range of late PIE so Proto Indo-Iranian wouldn't have had formed at that stage.

    ReplyDelete
  80. You take this suggested dating too literally. Common Indo Iranian if such a language existed could have been anywhere from 4000 bc to 1500 bc if one is forced to put a range. Nothing tells us that there ever was a PIE language with abrupt beginning at 4000 bc. The wheel line is not hard dating.. it’s a single word for gods sake!

    ReplyDelete
  81. This is a linguistic tree from Reich’s lecture:

    https://s7.postimg.org/jhiotcuop/screenshot_308.png

    This tree supports Andronovo –> India model.
    The question is when Indo-Iranian languages originated? Was it Sintastha or after an admixture from node 1 took place?

    ReplyDelete
  82. @postneo
    It's always funny when linguists try to put a floor on antiquity of languages by using reconstructions when they don't have an iota of evidence on existence or dating of the parent language.

    ReplyDelete
  83. @Chetan

    Indra lived in a fort and his Asura enemies also lived in a fort. This doesn't prove that Indo-Aryans were haters of urbanism. Remember Indra was called Asura a few times in the RV, so unless the IA worshipped the natives that's not true.

    ReplyDelete
  84. @Chetan

    You wrote: "I see one last possibility. That the movement 1 to B was pretty early- in the early stages of Yamna (3300-3000 BC), but in that case, we shouldn't expect the language to have been Proto Indo-Aryan or even Indo-Iranian at all. It would have been an archaic language similar to Tocharian".

    Interesting and possible, but even at 2500 BC this archaic (Late-Proto)-Indo-European could have arrived in South Asia. Anyway, aDNA samples taken in Rakhigarhi were not earlier than 2700/2600 BC, as the cemetery in mound 7 is from Mature Harappan phase (2600 to 1900 BC)or 2700 to 1900 BC if we take into account new calibrated C14 datings there.

    ReplyDelete
  85. @zulla
    The Rig Veda is definitely a bronze age text (As evidenced by various technologies it references), suggesting that it was composed in the neolithic or even mesolithic is ridiculous. The sanctity of a river doesn't necessarily have anything to do with its size. A narrow slow flowing rain fed river might have been far more suitable for settlements than a much larger glacier fed river prone to flooding.

    ReplyDelete
  86. @jijnasu

    We know that Rig Veda Mandalas are from different eras.

    Which technologies are we talking about?

    7000BC copper bangle and copper arrowhead found in pre Harappan Bhirrana?
    https://en.wikipedia.org/wiki/Bhirrana#Dating

    4000BC copper 6 spoked wheel like amulet in Mehrgarh?
    https://www.washingtonpost.com/news/speaking-of-science/wp/2016/11/15/this-6000-year-old-amulet-is-the-oldest-example-of-a-technology-still-used-by-nasa/?utm_term=.d54befbc4fa4

    Then there is the absence of the word 'istaka' (brick) in the Rig Veda, while it finds mention in the later texts like Yajurveda. If the geography of the RigVeda is Punjab/Haryana, how did this happen? Bhirrana pre Harappan 7000bc phase shows lack of bricks, and subterranean dwelling pits plastered with alluvium from Saraswati river. The use of mud bricks for settlements commences from the next phase ie early Harappan.

    Rig vedic people were definitely not 1500-2000bc.

    ReplyDelete
  87. Article in FT on Reich’s book:

    https://www.ft.com/content/824635cc-2c60-11e8-a34a-7e7563b0b0f4

    ReplyDelete
  88. @EastPole

    I can't access it. Does it say anything interesting about Central and South Asia?

    ReplyDelete
  89. No, nothing special, some quotes:

    https://s14.postimg.org/i3qquvl8h/screenshot_345.png

    https://s14.postimg.org/j60xdh97l/screenshot_346.png

    https://justpaste.it/1iogy

    ReplyDelete
  90. @eastpole

    so the overwhelming archaeological & now genetic evidence so far concludes that any male mediated 'invasion' post 2500bc occurred from Yamna into west europe.
    now waiting for south asian aDna.

    ReplyDelete
  91. @David

    The trick to read FT is paste the URL in google. If it turns up, it turns up with a readable link. It has some interesting things:

    "But the genetic evidence leads Reich to challenge the conventional view of one-way traffic out of Africa.

    He proposes an alternative scenario: the first migration of hominins (Homo erectus) into Eurasia 1.8m years ago led to substantial evolution of archaic humans there too. Some of them then moved back into north Africa, where they became the primary founders of the population that later evolved into modern humans. While this is unproven, Reich writes, “the evidence for many lineages and admixtures should have the effect of shaking our confidence in what to many people is now an unquestioned assumption that Africa has been the epicentre of all major events in human evolution.

    ReplyDelete
  92. This comment has been removed by the author.

    ReplyDelete
  93. @Al Bundy

    I think of that doubt on the African Origin as far, far more explosive than the whole India issue. Chris Stringer also has hinted on it before.
    The issue is the mtDNA of Sima de los Huesos. Either it means Denisovans admixted with that specific Proto-Neanderthals to give that band that mtDNA; or Anatomical Modern Humans were in the neighbourhood to replace their mtDNA early enough to completely replace it.

    Either way: That means the hotspot of the divergence was near Spain.

    ReplyDelete
  94. @Alogo

    Following your suggestion, I've re-labeled Greek:NA17376 as Greek_Crete:NA17376.

    But I removed Greek:NA17374 and Greek:S_Greek-2 because they're outliers, and I don't know for sure that they're Trabzon Greeks.

    ReplyDelete
  95. @zulla
    How could the rig veda have been composed millenia before the chariots and the wheel existed? Also while there might have been some time gap between the different mandalas and the rig veda and the other vedas, to argue that the rig veda was composed over 5000 or more years is ridiculous. It is not possible for language to stay unchanged for so long a period

    ReplyDelete
  96. @jijnasu

    But the later mandalas, especially 10th are surely in younger sanskrit.
    We will never know in what form the earliest mandalas were originally written/transmitted.

    Let me ask you this. If the 'aryan' charioteers of Sintashta invaded an almost dead/deserted IVC using their chariots, why did they suddenly stop burying the chariots in IVC with their dead elite like they did in Sintashta? also, why no fire altars excavated in Sintashta? after all agni is one of the most worshipped deities.

    enough evidence of toy carts, spoked wheel copper amulets, terracota wheels with painted spokes have been dug up in IVC. rig veda chariots were made from wood, which would not have survived till present given the climate of north india.

    ReplyDelete
  97. @zulla
    Entirely possible that verses were modernized but even then 5000 years is a huge stretch of time, the language would be entirely unrecogonizable. For comparison separated by a little nore than 2000 years ashokan prakrit differs hugely with hindi in both grammar and vocabulary

    ReplyDelete
    Replies
    1. Sry, posting from different account on my phone.

      Official/religious and public spoken languages have different rates of evolution. Already we are talking about 3500 years of preservation since 1500bc, so nothing is impossible and there is nothing to compare rig veda to.

      Also given the importance of pronunciation and sound in vedic sanskrit, I'm assuming it was easier to control the language through the shakha system.

      Delete
  98. @Salden

    Well, Reich just did. I am surprised how this is not discussed more.

    ReplyDelete
  99. @salden this africa business in the interview is good to sell books.

    Someone help me understand. @Matt had posted this months back.

    Cheers. So in this topology the "ANI" for populations is (ordered by steppe):

    Kalash - 55:45 Steppe:Iran_N
    Brahmin_India - 55:45 Steppe:Iran_N
    Gond - 49:51 Steppe:Iran_N
    Balochi - 35:65 Steppe:Iran_N

    Then you have (ordered by "ANI"):

    Balochi - 84:16 ANI:ASI
    Kalash - 81:19 ANI:ASI
    Brahmin_India - 64:36 ANI:ASI
    Gond - 22:78 ANI:ASI


    If Gond can be modeled as 49% steppe, 51% Iran_N and also as 22%ANI and 78%ASI; does this mean that both ANI and ASI can be modeled as Steppe & Iran_N in varying degrees?

    ReplyDelete
  100. @Salden

    Right.

    The way I read it, the "OoA" theory is more than a geographical pinpointing of the origin of humans. It is a theory that mankind evolved in Africa from Homo Erectus, via Homo Heidelbergensis to Homo Sapiens and then spread over the world, displacing earlier exoduses of both Homo Erectus and Homo Heidelbergensis.

    But I don't care about the semantics, I want to know what the genetic evidence is David Reich mentions. The mtDNA of Sima de los Huesos? Or maybe even more?

    I could think of the fact that genetics points to a separation of H.Sapiens, Neanderthals and Denisovans round 700.000 ya, but we only have Central African samples of less than 200.000 ya. The only older samples we have are Morrocan.

    Anyway, I'd say a highly interesting point.

    ReplyDelete
  101. @Salden

    " but we only have Central African samples of less than 200.000 ya. "

    I forgot that there is also the South African Florisbad skull: 260 kya. But one can possibly argue that H. Sapiens can't have been in South-Africa by the time Sima de los Huesos lived because H. Naledi lived there at that time. And we know from other examples - Neanderthal, Flores man, Denisova, that archaics never lived long alongside H. Sapiens.

    ReplyDelete
  102. @zulla, my comment meant, in that model, ratios of Steppe:Iran_N *within* the "ANI" for different populations. That is that Kalash's 81 ANI would break down into 45 Steppe + 36 Iran_N (following that ratio), while Balochi's 84 ANI would break down into 30 Steppe + 54 Iran_N, etc. The ratios there aren't alternative models. Can't remember much more about that model.

    ReplyDelete
  103. About the Aryan migration

    http://www.currentscience.ac.in/Volumes/114/05/0971.pdf

    ReplyDelete
  104. @chetan
    Absolutely no content in the paper in terms of new samples... Written like a school essay, discussing low resolution clades and 10 year old examples. Why bother publishing such a paper in 2018?

    ReplyDelete
  105. This comment has been removed by the author.

    ReplyDelete
  106. @postneo I knew it isn't much. Just posted the link here in the hope there may be something in it about the upcoming papers.

    ReplyDelete
  107. @chetan, a possibly interesting sentence from that paper (by Prof. Majumder), assuming he knows something about upcoming indian aDNA paper, is "The most likely period of entry of [R1a1a] into India is between 4500 and 5000 years ago. " Unless he is referring to U2i.

    ReplyDelete
  108. @HistoricallyCurious

    This is extremely interesting! Thanks for bringing that up here. According to the English in the article, the author is very much referring to R1a1a as you suggested. So since the author put a very tight likely period of entry as between 3000 BC and 2500 BC and not later, are we to think that the findings could support a hypothesis that the urban civilisation of the Indus Valley was established by incoming Indo-Aryans from somewhere north and this event may have triggered Dravidian migrations to south India where they first diversified into NDr., CDr., and SDr.?

    ReplyDelete
  109. @epoch213

    How on earth would archaic human stuff from hundreds of thousands of years ago possibly be more explosive than information bearing on the founding of a major world religious tradition, interpretation of scriptures, post-colonial politics, and the ethnolinguistic identity of zillions? Do you figure there's frickin' aliens involved or something?

    ReplyDelete
  110. Is that node A going from Central Anatolia to South Caucasus?

    ReplyDelete

Read the rules before posting.

Comments by people with the nick "Unknown" are no longer allowed.

See also...


New rules for comments

Banned commentators list