Thursday, August 13, 2015

Kets are rich in Ancient North Eurasian (ANE) ancestry


There's an interesting and very thorough preprint at bioRxiv looking at the genomic structure of Kets, the last nomadic hunter-gatherers of Siberia. From the paper:

Based on all analyses, we can tentatively model Kets as a two-way mixture of East Asians and ANE. Therefore, ANE ancestry in Kets can be estimated using various f4-ratios from 27% to 62% (depending on the dataset and reference populations), vs. 2% in Nganasans, 30 ‒ 39% in Karitiana, and 23 ‒ 28% in Mayans (Suppl. file S7, see details in Suppl. Information Section 8). Integrating data by different methods, we conservatively estimate that Kets have the highest degree of ANE ancestry among all investigated modern Eurasian populations west of Chukotka and Kamchatka. We speculate that ANE ancestry in Kets was acquired in the Altai region, where the Bronze Age Okunevo culture was located, with a surprisingly close genetic proximity to Mal'ta. Later, Yeniseian-speaking people occupied this region until the 16th-18th centuries. We suggest that Mal'ta ancestry was later introduced into Uralic-speaking Selkups, starting to mix with Kets extensively in the 17-18th centuries.

I'd say these findings make a lot of sense. Below is a spatial map put together by Sergey, based on my K8 model, showing the distribution of ANE across much of Eurasia. Note the ANE peak of around 28% among the Kets.


Citation...

Flegontov et al., Genomic study of the Ket: a Paleo-Eskimo-related ethnic group with significant ancient North Eurasian ancestry, bioRxiv, Posted August 13, 2015, doi: https://dx.doi.org/10.1101/024554

200 comments:

  1. Expected, What are the dominant Hg's of the Kets?.

    ReplyDelete
  2. "Q
    95%
    Speaks of recent founder effect"

    Or drift. They have had a very small population for quite a while.

    ReplyDelete
  3. The ANE estimates fluctuate a bit due to the team having included Siberians as East Asian references in their models. Nganasans have only 2% ANE here because Yakuts were used as their East Asian reference.

    ReplyDelete
  4. Looking at that map, I think if we had some Paleolithic aDNA from Kazakhstan it will be high in ANE.

    ReplyDelete
  5. While the results are not new or surprising, the conclusions from them might be a bit misleading.

    First it should be noted that by ANE they mean very specifically MA1 admixture/affinity. Then the formal stats to measure it have some caveats. For example, something like:

    f4(WHG, Pathan; MA1, Chimp)
    f4(WHG, Lezgin; MA1, Chimp)

    might give the wrong impression that WHG have more ANE than Pathans of Lezgins (I don't have those stats at hand, but they're likely positive).

    Finally, saying that Kets have the highest degree of ANE ancestry among all investigated modern Eurasian populations west of Chukotka and Kamchatka when they estimate that they have from 27% to 62% sounds a bit sloppy (if they can't get a more accurate figure, better to omit it altogether).

    I think the conclusion should be something like "Kets have the highest MA1 related ancestry among modern Eurasians west of Chukotka and Kamchatka", and omit those figures would have made a better impression.

    The IBS sharing of MA1 made by David some time ago (top 25):
    Karitiana 0.674521
    Mixe 0.673172
    Pima 0.672787
    Surui 0.672689
    Mixtec 0.672055
    Bolivian 0.671739
    Mayan 0.671448
    Piapoco 0.671418
    Quechua 0.671086
    East_Greenlander 0.667868
    West_Greenlander 0.667753
    Eskimo 0.667554
    Ket 0.667259
    Mansi 0.666903
    Mari 0.666430
    Shors 0.666378
    Chukchi 0.666163
    Selkup 0.666092
    Itelmen 0.665718
    Erzya 0.665503
    Koryak 0.665259
    Estonian 0.665248
    Kalash 0.665190
    Lithuanian 0.665091
    Russian_Kargopol 0.665043

    ReplyDelete
  6. Karl_K,

    A founder effect is a subset of genetic drift.

    Genetic drift, in this context, is defined by a loss of genetic diversity over time. Founder effects = subset of original population expanding somewhere new (e.g. different Pamiri villages in Tajikistan). Bottlenecks = sudden reduction in original population due to external factors (e.g. Ashkenazi Jews due to the Holocaust).

    Whether it's a founder or bottleneck effect depends on Ket history. I don't know whether or not they suffered any significant depopulation events as a consequence of the Russian expansion.

    ReplyDelete
  7. @DMXX

    I believe that they were historically known to be in small hunter gatherer populations. They were also moved around and split up by the Russians.

    But for at least their last several generations they have had a population under 2000 people, and the surviving group may be from only one or two original populations, so yes, a founder effect is also likely.

    ReplyDelete
    Replies
    1. Yes but there is not reason to assume a priori that they are the same gatherer groups as one which dwelt in the same region 8 thousand years ago. In fact, Id bet they not

      Just look at other Y lineage founder effects in Siberia- R1b and N.

      If a continuous steady state population, it should have homogenised more by now

      Delete
  8. Obviously a major point of interest in this population is that their language has been hypothesized to be similar to some North American native languages.

    It seems quite reasonable that a later wave of Siberians, more closely related culturally and genetically to the Ket, made their way over to the Americas, but never made it all the way to South America.

    The Ket were said to have previously had a strongly patriarchial society. This seems like a trend in ANE dominant populations. And this could be the source of these languages in the Americas.

    Just as a side thought, I know that a large portion of the North East Native American population has R1 Y haplogroups, but most sources just say that that is probably from Europeans (which of course makes sense). But has this really been sorted out? Or has it been hindered by these groups not wanting to be tested?

    ReplyDelete
  9. Re: Link to NA languages, yes according to wikipedia:
    "In February 2008, linguist Edward Vajda also submitted a paper on the proposed link between Ket with the Na-Dene languages. Now published in 2010, Vajda's paper has been favorably reviewed by several experts on Na-Dene and Yeniseian languages, including Michael Krauss, Jeff Leer, James Kari, and Heinrich Werner, as well as a number of other well-known linguists, including Bernard Comrie, Johanna Nichols, Victor Golla, Michael Fortescue, and Eric Hamp, so that a broad consensus has formed in support of this connection[citation needed]. Some experts on Yeniseian remain extremely skeptical or reject the hypothesis (e.g. Stefan Georg)."

    ReplyDelete
  10. More info:
    https://en.wikipedia.org/wiki/Den%C3%A9%E2%80%93Yeniseian_languages

    ReplyDelete
  11. Yeah, Kets are very high in ANE, that's something we have known for some time already, so this is welcomed but not really big news. Basically Kets are the "purest" descendants of Ma1.

    But re. your K8 model, my main problem with it is that by ignoring the "Gedrosian" (or Baloch-Caucasian or Highland West Asian/ANI) component, as well as the ASI component, you get results that exaggerate the amount of ANE in West, South Asian and probably even European populations, as anything that does not fit with the other Eurasian components (in this case notably ENF/Arabian and will be split among them following affinity and that's why you get so high ANE levels in West Asia (and in other exercises also in South Asia).

    I tracked your K8 (ANE+K7) model to this entry like because your provided link is not informative enough.

    I'd suggest that, in order to get accurate results, you begin using some sort of K8 model, in which Southern and West Asian key components are factored. Else the results can't be taken seriously for that region.

    ReplyDelete
  12. Erratum: "you begin using some sort of K8 model" should read: "you begin using some sort of K10 model".

    ReplyDelete
  13. IMO R originated in SC Asia or Nearby so having high ANE affinity is natural, you can argue whatever you want but we need aDNA.

    ReplyDelete
  14. David, is there any genome sequencing available of Khotons?.

    ReplyDelete
  15. Maju,

    The Med/Levant vs West Asian dichotomy in the Near East is an illusion created by ANE. See here...

    http://eurogenes.blogspot.com.au/2014/12/ane-is-primary-cause-of-west-to-east.html

    Nirjhar,

    I don't have any Khotons, but I do usually run the Khakas from the Altai who show a high level of ANE.

    ReplyDelete
  16. Khotons have high levels of R1a, so it would be nice to see their ANE levels:)....

    ReplyDelete
  17. @David: I strongly disagree: the "teal component" is not any illusion. Anyhow you're also missing ASI, so I must insist that you are artficially inflating the ANE in all that region.

    ReplyDelete
    Replies
    1. I saw on anthrogenica someone had isolated the Geodrosia components from ancient samples. This greatly reduced the "steppe" component in both modern Central Asians and Europeans, which was about half that of Geodrosia (on average).

      Of course this has been subject to debate- and can only be disproven or proven with pre-Bronze age genomes from central asia

      Delete
  18. Interesting paper. Perhaps Q and R groups from Steppe really did spread some/type of ANE type component.
    Lezgins/Tabassarans should have near 0% ANE if the "teal component" found in ancient Yamnaya is solid, since both of these groups have predominantly 1 ydna[excluding R&Q], while Yamnaya males had a variety of females, some perhaps from Near East[non U4 or U5].
    Instead; some of the highest numbers of ANE in K8, in and around the Caucasus, from these two groups.

    https://en.wikipedia.org/wiki/Y-DNA_haplogroups_by_ethnic_group

    ReplyDelete
  19. The real problem is the lack of aDNA from Asian Sites....

    ReplyDelete
  20. We do have aDNA from the region... it's the same reference used here for ANE - Ma'alta. He had 0% East Asian affinities, so this is obviously a more recent intrusion in the Kets.

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

    ReplyDelete
  22. According to the map, ANE drops in Ucrania and Greece? Interesting

    ReplyDelete
  23. Oops, it's not Ucrania but Moldova - I should wear my glasses more often

    ReplyDelete
  24. Oops, it's not Ucrania but Moldova - I should wear my glasses more often

    No, they are Ashkenazi Jews, not Moldovans. I do not know why but Ashkenazi Jews are located on Moldova on that map. You should certainly wear your glasses more often. :)

    ReplyDelete
  25. @kArl_K,
    "Just as a side thought, I know that a large portion of the North East Native American population has R1 Y haplogroups, but most sources just say that that is probably from Europeans (which of course makes sense)."

    If basal forms of R1 exist in Native Americans that'll give plenty of room for R1a-M417 and R1b-M269, and their ancestors, to have developed North of the Caucasus.

    ReplyDelete
  26. @Cromagnon: "I saw on anthrogenica someone had isolated the Geodrosia components from ancient samples. This greatly reduced the "steppe" component in both modern Central Asians and Europeans, which was about half that of Geodrosia (on average)".

    Maybe. I recall some study (but can't recall where or which one, sorry, maybe at Razib's?) that analyzed ANI for European ancestry and this was just a very small fraction, suggesting that something like 90% of it was of direct West Asian origin and hence Neolithic, rather than steppe-IE.

    ReplyDelete
  27. You can't use modern Europeans as a proxy for ANI, because modern Europeans are only partly of steppe origin, and South Asians also have extra West Asian Neolithic input.

    So the overall ANI effect looks Georgian in formal stats, and Gedrosia in Admixture, when in fact ANI is in large part of steppe origin.

    Let's all raise our hands in the air and shout Gedrosia, Gedrosia, Gedrosia. Do you feel better now?

    ReplyDelete
  28. Davidski,

    This may be slightly off topic, but would you please run Kennewick Man's DNA through your K8 calculator? I'm sure a lot of people would be interested in his ANE, WHG, and other proportions. A sample has been uploaded to GEDmatch as kit F999970. And a fuller sample is linked to from here: http://www.y-str.org/2015/07/8300-year-old-ancient-dna-of-kennewick.html

    ReplyDelete
  29. Regardless, David, check please the Raghavan 2013 data and ponder please why the shocking difference in results between them and your K8. To me it is obvious: you are forcing West Asians and South Asians to fit either with Europeans (WHG, EEF) or with East Asians or with Ma1, so they artificially appear much more ANE than they actually are (that's a problem with "zombies" when they are poorly chosen). Per the Raghavan data they are at Sardinian or at best Southern European levels of ANE affinity, in you data they are generally higher than most Europeans. Your results are wrong because they lack West and South Asian "zombies" that can serve as contrast with ANE or, alternatively, because you didn't use the effective simplicity of "f3(Yoruba; MA-1, X)".

    ReplyDelete
  30. @Maju,
    "Per the Raghavan data they are at Sardinian or at best Southern European levels of ANE affinity, in you data they are generally higher than most Europeans."

    "WHG" in Europeans raises their relation to MA1 independent of "ANE" or "EHG" or "Teal" ancestry. MA1 was closely related to WHG.

    ReplyDelete
  31. You should know full well that WHG increases affinity to ANE, while SSA and ASI dampen it. That's why Europeans show higher affinity to MA1 than West and South Asians do, despite generally having lower levels of ANE than them.

    Maju, it's obvious that at this point you're out or your depth here. Please do a lot more reading before continuing, because you're simply not making any sense.

    ReplyDelete
  32. "(e.g. Ashkenazi Jews due to the Holocaust)."

    Wasnt the Ashkenazi bottleneck calculated to be in the middle ages, a few years ago?
    Wich, I forgot the actual number, but something like "effective population size of 400" or something? I can somehow recall some craisy claim like "All Ashkenazi Jews that live today stem from 400 Jews that lived 800 years ago."

    ReplyDelete
    Replies
    1. Not a bottleneck but a founder effect in late roman period, wasn't it ?

      Delete
  33. It looks like Mesolithic Siberia is an affair for yDNA Q and R and probably C-M217 (C3) and a mysterious paleo-Melanesian line if it is ever discovered.

    This paper has answered many of the questions I raised in my previous posts.

    Motala: in their tree Fig. 2, C, there is an admixture edge from a Ket - Malta like population into Motala. According to the paper, ANE ancestry in Motala12 was estimated at 22%. Now we know that Karelian hunter gatherer was R1a and C1, and among Motala samples there was one sample that was close to being assigned to hg Q. IMO, until any other evidence is obtained, this should be the answer why Motala carry EDAR and have traces of ENA.

    ANE ancestry in Arctic people: the paper notes that “according to a single f4 statistic, the Kets had the third highest value of ANE genetic contribution among all Siberian ethnic groups, preceded only by Chukchi and Koryaks.” According to the IBS sharing of MA1 made by David some time ago (top 25), Eskimos have even slightly more MA1 related ancestry than Chukchi and Koryaks. In the admixture runs, these populations are completely East Asian but have thus 67% “West Eurasian” ANE. The new paper also concludes that Kets and Selkups belong to a group of populations most closely related to ancient Paleo-Eskimos in Siberia, and we know that Saqqaq was Q-MEH2. He belonged to Arctic Small Tool tradition bringing pottery, bow and arrow technology to the northern North America.

    Is Ma1 mixed or not: the paper and its supplementary material contains a myriad of trees, so you can find support for many claims. Ma1 is usually placed on the West Eurasian branch and sometimes on the East Eurasian branch but it is nearly always near the root of the East-West divide. In Fig. 2, A. there is an admixture edge from Ma1 to Chipewyan and to Eskimos. In trees 9.5. A. and 9.14. A., there is an admixture edge going from Central and South Americans into Ma1 and in 9.13. A., 9.15. A., 9.16. A., 9.17. A., 9.19. A., 9.20. A. and 9.21. A. from Karitiana into Ma1. In the end, admixture probable happened in both ways. When Ma1’s population formed, it mixed with a population that was at Ma1’s time already in Beringia, but later on also a Ma1-related population made it to America.

    Finnish ENA: this paper also gives an explanation to Finnish ENA. It notes that “the 'Ket' component also occurred at a low level in Russians (up to 7.1%) and Finns (up to 5.4%)”. Admixture run Fig. 1, A, shows that Finnish ENA is Ket-like and not Nganasan-like, and Finns may even proportionally have more Native American than Kets. I could imagine that a smaller part of it is from Butovo (i.e. Karelian hunter gatherer) and the rest from Comb ceramic complex. I hope that we will one day obtain yDNA results from Comb ceramic complex.

    Uralic groups: in Fig. 2, A., there is ad admixture edge going from Ket into Uralic speaking Mansi which should mean that there would be a Ket-related substrate in Ob-Ugric languages. One Q-M242 was also found in 60 Mansi males. Maris appear in most admixture trees. They are always on the West Eurasian branch of the tree. There is often an admixture edge going from Late Dorset branch or Ket - Late Dorset branch or only from Kets into Maris. There is sometimes also an admixture edge from Maris to Kets. Maris have 14.3 % of Ket component. Their ENA portion is c. 30%. As it is said that Maris have a portion of Turkic ancestry, we could assume that ENA in Mari is for the most part Ket/paleo-Siberian and Turkic related.

    Linguistic conclusion: after reading this paper, I am more convinced that Uralic languages formed during the Bronze Age/Iron Age upheavals from a language that was close to Proto-IE and has several paleo-Siberian languages as substrate languages.

    ReplyDelete
  34. @David: "WHG increases affinity to ANE".

    No. It's possible that (in X circumstances) WHG shows a correlation with ANE. That would make some sense. But if there is a fixed WHG component and a fixed ANE component by design, one detracts from the other, so in a hypothetical K=2, greater WHG = less ANE and vice versa (there's no other option).

    "Maju, it's obvious that at this point you're out or your depth here".

    No. It's pretty obvious that I caught your pet theory in a major contradiction. No need to go defensive, you can surrender to the evidence because it's a clear case of Q.E.D.

    Honestly, I don't understand how so much ado was made about ANE without really considering seriously the original Ma1 paper and the measures it already provided.


    @Krefter: ""WHG" in Europeans raises their relation to MA1 independent of "ANE" or "EHG" or "Teal" ancestry. MA1 was closely related to WHG".

    Ma1 and ANE are the same thing, unless someone has redefined it and hasn't told the public. Anyways, excluding BAE admixture, WHG and EEF (Lochsbour and Stuttgart) are clearly closer to each other than to ANE (Ma1). For example: Lazaridis' model tree (but there are others).

    Lazaridis did not consider ANE to be "Western Eurasian". For me however all three are Western Eurasian just that (pre-)EEF and (pre-)WHG shared more paleohistory (exchanged more genes but not the "BEA" thing, which must be a late incorporation) while ANE was more isolated in its Siberian backwater corner. However it should have incorporated some Paleoeuropean genetics from Gravettian, what may make it closer to WHG (which may well be 100% descended from those roots) than to EEF (which is only c. 50% Paleoeuropean). This is difficult to gauge, so it'd be more useful to discuss on actual specific data, I'm sure you have something in mind but I can't read your mind.

    But well, all this is beyond the point, what is clear is that Ma1=ANE and that, judging with the proper approach (i.e. like Raghavan did), ANE is very low in all West and South Asia, being something characteristic of North Eurasia (as the name suggests) and even more intensely of Native Americans.

    ReplyDelete
  35. WHG increases affinity to MA1 (ANE) because the more WHG the less Basal Eurasian ancestry.

    Basal Eurasian is highly diverged from ANE, therefore it dampens affinity to MA1.

    WHG is much less diverged from ANE, therefore it can only dampen affinity to MA1 if we're comparing samples that are mixtures of WHG and ANE, as well as closely related components like EHG.

    As soon as Basal Eurasian, Sub-Saharan and other highly diverged components come into play you'll start seeing a lot of discordance between affinity and admixtures levels.

    Maju, you really need to update yourself.

    ReplyDelete
  36. Yes, "BEA" (whatever it is) clearly dampens Ma1 or whatever other "true Eurasian" affinity. I don't see how this affects India in any way. The most affected would be Palestinians, Bedouins and other related SW Asian populations but not South Asians.

    I was just reading (by recommendation of one of your supporters) the Ayub & Mezzavilla paper on the Kalash and it generally confirms my understanding of the matter:

    · Fig. 3a shows Kalash as unusually strong in ("absolute") Ma1 affinity for the region but West Asians as lower in ANE than Sardinians and other Pakistanis lower than all non-Sardinian Europeans. The method is the same as Raghavan's (f3 with Yoruba as control).

    · Fig 3b (left) shows an "admixture triangle" (French, Han, Ma1) that is comparable in concept and results to your K8 scheme and there the results are much different, exaggerating the ANE affinity of Pakistanis. This is clearly because of lack of local "pre-ANE" references. Following Raghavan I'd suggest to use Iranians and Malayam (or Paniya) because these populations are as low in "absolute ANE" as Sardinians in Europe.

    Kalash are strong in Ma1 but their coalescence as a distinctive isolate population is determined to be some 12,000 years old, what should tell us something about the arrival of ANE to the region. So nothing to do with Kurgans, Indoeuropeans, etc.

    ReplyDelete
  37. @Maju

    Now I see your point. But I see some problems there. And I guess it all comes to not having a more modern sample of ANE than MA1.

    MA1 indeed gave us a decisive clue about ANE. But to say that ANE = MA1 is a bit like saying that ENA = Ust'-Ishim, or that WHG = Kostenki-14 (If we didn't have any "real" WHG sample, we'd have to do with Kostenki as the closest thing, but I guess you see how problematic that would be).

    So yes, maybe the synthetic ANE used in K8 is not perfect, though I think it's quite good. I've only observed (cross checking with other admixture runs and IBS based PCA plots) that it underestimates ANE in Near Eastern populations and to a lesser degree south European ones.

    The f3(Yoruba; MA-1, X) kid of stat has the problem that it's a bit sensitive to (in this specific case) BEA and ASI admixture is the X populations. It's not nearly as sensitive as D-stats, but still quite more sensitive than IBS sharing (Matt actually kindly made a graph showing this, I'll look for the link). Admixture should be the best option, but as I said, we lack really good ancient samples (say, 8000 years old, instead of 24 Kyo).

    In any case, the MA1 IBS sharing I posted above clearly show that Kalash share almost as many alleles as Kets with MA1. So I'm not sure why you think that ANE is low in S-C Asia. With good samples and with Admixture, it will only be confirmed that it actually peaks there.

    Also I wouldn't give much importance to the fact that MA1 was found in Siberia. The good modern samples will probably come from near NE Iran/Afghanistan, where you have well argued for the origin of R1 (though maybe around the Pamir mountains will be a better place to look, since Neolithic -and ENF with it- arrived there quite later and the climate might have preserved better the samples).

    ReplyDelete
  38. South Asians have significant Basal Eurasian admixture, most of which is probably from Neolithic Iran.

    Also, there seems to be something strange about ASI, because it gives Sub-Saharan admixture to South Asians in many Admixture runs. This might also be responsible for dampening affinity to MA1.

    ReplyDelete
  39. Here is the link:

    http://i.imgur.com/myboZPF.png

    Basically it compares the affinity to MA1 using the 2 methods: IBS sharing and f3 stats with an outgroup. As you can see, the affinity to MA1 is reduced clearly by the ENA admixture (and SSA) in f3 stats when compared to IBS sharing. IOW, European populations (who lack ENA and SSA) show higher relative affinity (vs. Asians) to MA1 when using f3 stats (vs. using IBS sharing).

    So for this cases IBS sharing is a better method (though not as good as a good Admixture run).

    ReplyDelete
  40. Re: Kalash, they're basically very similar to Pathans and other neighbours and we shouldn't expect any significant difference. The complete list of IBS sharing puts Tajiks and Pathans just marginally below Kalash.

    ReplyDelete
  41. @Alberto: I am not the one who has defined what is the reference for ANE. ANE = Ma1, exactly the same as WHG = Lochsbour or EEF = Stuttgart. The definitions probably date to Lazaridis 2014 (which was pre-published and widely discussed already in 2013, soon after the Raghavan paper). There is another somewhat more recent ANE reference which is Afontova Gora, but nobody seems to be using it. Otherwise there are just no references. ANE is NOT a recent (Holocene) component but a Pleistocene one (unlike WHG and EEF). It seems to be important but it is not directly comparable to other components defined on Holocene aDNA, and therefore more directly impacting modern genetics. This makes ANE a bit of a ghost, especially if we insist on claiming it only associated to Indoeuropean expansion, what is not the case very apparently in many cases (to test IE expansion demic impact is much better to use Yamna or CW or Sintashta references, logically).

    "So yes, maybe the synthetic ANE used in K8 is not perfect, though I think it's quite good".

    No, it is not "good" at all: it is flawed in its very concept. I know it sounds very harsh to the work of David, but that's what it is. Someone has to evidence the reality of the "naked emperor", I'm just good at that.

    "I've only observed (cross checking with other admixture runs and IBS based PCA plots) that it underestimates ANE in Near Eastern populations and to a lesser degree south European ones".

    I would not consider PCA too useful here (PCAs have only limited usefulness, roughly like a K=3 or K=4 ADMIXTURE plot) but you got me intrigued re IBS (identical by state). However not all IBS is IBD (identical by descent), so I'd like more info on that before I can judge.

    "MA1 IBS sharing I posted above"...

    OK, I'll read your previous comments to see if I can find the link.

    "The f3(Yoruba; MA-1, X) kid of stat has the problem that it's a bit sensitive to (in this specific case) BEA and ASI admixture is the X populations".

    BEA yes, ASI not. ASI is as non-African as any other Eurasian component (East Asian, WHG, etc.) There is no problem there. In fact the problem stems largely for not using ASI (or equivalent reference), what exaggerates the ANE results. Also "Gedrosia" (ANI), which is quite apparently mostly non-ANE and non-European but a bit harder to fine-tune than ASI (it should not be ignored in any case).

    "The good modern samples [of ANE] will probably come from near NE Iran/Afghanistan"...

    No way. Per Raghavan Iranians are very low in ANE, similar to Sardinians. The impact or ANE is probably greater in Pakistan and North India (and surely much more related to Kalash late Pleistocene coalescence rather than IE migrations, but lower than in non-Sardinian Europe in any case) so the reference in the subcontinent, if we want to align all regions with a Sardinian-like baseline, should be something like Malayam or Paniya (i.e. roughly close to pure ASI).

    ReplyDelete
  42. OK, Alberto, I just found your comment on IBS. The only noticeable item are Kalash, which, as Ayub & Mezzavilla note are strikingly high in ANE for their region but also an ancient genetic isolate dating, they claim, to 12 Ka ago. Otherwise I see nothing and I have already accounted for the Kalash exceptionality.

    ReplyDelete
  43. @Maju,

    MA1 had some-type of common ancestry with Kostenki-14 and WHG. Higher WHG=added affinity to MA1.

    ReplyDelete
  44. Maju,

    Figure 3b is nothing like David's K8, which is an ADMIXTURE run. Figure 3b is about shared drift, just like figure 3a. But, figure 3b is an important expansion of figure 3a, as it involves the plotting together of different f3 stats to facilitate comparison and correlation. All the Pakistani populations clearly exceed Europeans in MA1-affinity, as per figure 3b. This is in the context of MA1 vs French, so Europeans are much closer to the French than they are to MA1, while Pakistanis are more shifted towards MA1 in this context. Basically, Pakistanis almost have the same shift towards MA1 versus French as Native Americans have towards MA1 versus the Han (and the Kalash are even more extreme than Native Americans, in the context of West Eurasian MA1-affinity, and when Native Americans are compared to Han)!

    Again, the f3 stats aren't in anyway measures of ANE admixture, but measures of ANE affinity, which is affected by many different factors. As per Lazaridis et al., Lezgians and Chechens have more ANE ancestry than any population in Europe. If we continue to hold the ANE-WHG model, West and South Asians do have more ANE than Europeans, and South Asians tend to have the most actual ANE admixture in all of Eurasia.

    Though, new aDNA does change things, but that is a different story.

    Regardless, again, I don't think your'e being very coherent right now. All you need to do is try some reading on the methods, and you'll be up to date. I mean no disrespect, I'm just telling you this because the discussion will make more sense to you if you do so.

    ReplyDelete
  45. "South Asians have significant Basal Eurasian admixture, most of which is probably from Neolithic Iran".

    Where do you get that from? In my understanding BEA only seems to be present in EEF and some West Asian populations (south and west of Iran, not really much in Iranians) but I don't think there have been significant studies on this BEA mystery thing to be judgmental.

    In any case how is that different from BEA in Europeans or your reconstructed ENF (proto-EEF) zombie?

    "Also, there seems to be something strange about ASI, because it gives Sub-Saharan admixture to South Asians in many Admixture runs".

    Never seen anything like that. If you exclude all local references, South Asians, as any other, will have to split their affinities within the allowed "zombies" (pre-determined references), so it's possible that they will tend somewhat to Africa but this is even more notorious in Europeans and West Asians when you produce a K=2, for example, in which East Asians and Tropical Africans are the two references. This kind of fallacious analysis has led to many errors in assessment, beginning with the infamous claim by Cavalli-Sforza of Europeans being 2/3 Asians and 1/3 Africans, that we now know is radically wrong, but is still now and then wavered by some "hyper-Africanists" in order to exaggerate the African fraction of ancestry not only in West Eurasians but also in admixed Africans from the North and NE. I think this is the same issue mutatis mutandi.

    ReplyDelete
  46. Basal Eurasian is a major component throughout West Eurasia today, and especially in the Middle East. This should be obvious to anyone with even a faint interest in this area.

    This discussion is getting stupid.

    ReplyDelete
  47. @Maju

    I surely didn't mean to blame you for the definition of ANE. I know it is defined as MA1 in the original paper. But I just think that a 24 Kyo sample is not good enough. I put the example of Kosteki because if we lacked those other WHG samples (loschbour, KO1, La-Brana) we would be in a similar situation, and getting wonky results. We are just not lucky enough to have more modern samples of ANE.

    (Of course, here you can argue that the presumed modern samples should be given a different name, and I can agree with that. ANE can stay as MA1 for all I care. But it will stop being very useful, as Kostenki is not very useful today).

    As for where we will find good modern samples of ANE, we'll just have to wait and see. I can't argue with hard facts for it.

    Here a longer list of IBS sharing that includes Tajiks and Pathans:
    Karitiana 0.674521
    Mixe 0.673172
    Pima 0.672787
    Surui 0.672689
    Mixtec 0.672055
    Bolivian 0.671739
    Mayan 0.671448
    Piapoco 0.671418
    Quechua 0.671086
    East_Greenlander 0.667868
    West_Greenlander 0.667753
    Eskimo 0.667554
    Ket 0.667259
    Mansi 0.666903
    Mari 0.666430
    Shors 0.666378
    Chukchi 0.666163
    Selkup 0.666092
    Itelmen 0.665718
    Erzya 0.665503
    Koryak 0.665259
    Estonian 0.665248
    Kalash 0.665190
    Lithuanian 0.665091
    Russian_Kargopol 0.665043
    Chuvash 0.664942
    Finnish 0.664814
    Tajik_Pomiri 0.664787
    Moksha 0.664772
    GujaratiA 0.664268
    Russian 0.664208
    Pathan 0.664024
    Polish 0.663917
    Brahmin_UP 0.663913

    ReplyDelete
  48. @Alberto: the second graph is very interesting, yes. However, barring a few "extreme" populations it shows a strong correlation between Ma1-IBS and "absolute" Ma1 affinity (f3). The extreme populations tend to have stronger IBS than "absolute" Ma1 affinity, suggesting that some factors are masking their true Ma1 affinity. However they fall in three very specific categories:

    1. Native Americans
    2. Australo-Melanesians (probably because Denisovan admixture pulls unduly, but logically, towards Africa, towards the root)
    3. North Africans

    Even Naqab Bedouins are very close to the norm, so, barring these three cases, I see no reason to claim strong disconnection between "absolute" (f3, Yoruba-controlled) and plausible (IBS-inferred) Ma1 affinity.

    "IOW, European populations (who lack ENA and SSA) show higher relative affinity (vs. Asians) to MA1 when using f3 stats (vs. using IBS sharing)".

    Only very slightly. I would consider them within the norm. In any case it suggests that Ma1 "absolute" affinity is slightly exaggerated in Europeans, probably because Ma1 was already a bit Paleoeuropean (Gravettian influence, just a hunch anyhow). But the distortion is minor anyhow.

    One could indeed use that graph to determine correction parameters for each population but for the main cluster along the red line, these would only change things very little.

    ReplyDelete
  49. "I surely didn't mean to blame you for the definition of ANE. I know it is defined as MA1 in the original paper. But I just think that a 24 Kyo sample is not good enough."

    17 Ka if you use Afontova Gora. Anyhow, if you don't use these, then it will not be ANE but something else.

    The results you post are not substantially different (if Europeans are corrected down a bit, as your graph suggests that should be done) to what Raghavan et al. shows in fig. SI-21. The overall pattern is the same: Kalash first, then Tajiks and Burusho, then Gujarati, then Sindhi, Brahui, Balochi, etc. What varies is the correlation with Europeans, what I guess is justified per your linked graph (it may seem dramatic but it is actually just a small jump of about 0.005 in the Raghavan scale). Following up Iranians should still be near Sardinians, as they are actually quite lower in Raghavan's data, so IMO Iranians and, say, Paniya are good regional "Sardinian-like" baseline references.

    ReplyDelete
  50. @Sein: I know the figure is not the same but the errors introduced by lack of local references are very much the same, the resulting distortion, brutal exaggeration of Ma1 is roughly the same. The graph is therefore useless, as is the K8.

    "All the Pakistani populations clearly exceed Europeans in MA1-affinity, as per figure 3b".

    Junk in: junk out. That exercise is as useless or as misleading (if you happen to take it seriously) as K8. You can use Europeans and East Asians as controls for West/South Asia: it's just absolutely wrong!

    "so Europeans are much closer to the French than they are to MA1"

    LOL. And French, if plotted, would be 100% French, 0% Ma1. French are Europeans and other Europeans are largely like French.

    That does not happen with South Asians. For South Asians you have to use South Asian references (I suggest Iranians and Paniya, which are very low in "raw ANE" and should produce a baseline directly comparable to Sardinians in Europe).

    ReplyDelete
  51. Erratum: "You can use Europeans and East Asians as controls for West/South Asia: it's just absolutely wrong!" should read "You cannot use Europeans and East Asians as controls for West/South Asia: it's just absolutely wrong!"

    ReplyDelete
  52. Maju,


    In that case, it is absolutely wrong to use Han and MA1 as controls for Native Americans. Native Americans are far more diverged from East Asians than South Central Asians (Pamiri Tajiks, Tajiks, Kalash, Pashtuns, etc) are from the French or any other Europeans, and the true source of West Eurasian-affiliated ancestry in indigenous Americans is far closer to the EHG samples from Haak et al., rather than to MA1.

    Also, it is very strange to claim that West and South Asians don't have higher ANE ancestry than Europeans based on shared drift. As noted in the supplementary data of Lazaridis et al., although most Europeans share more drift with MA1 than NE Caucasians, NE Caucasians have higher ANE admixture than any European population. The discordance between ANE affinity and actual ANE admixture is then explained in that same supplementary note. Please refer to that part of the paper's supplementary data.

    ReplyDelete
  53. Maju, several TreeMix trees of this recent paper show an admixture edge from Dinka to Stuttgart. It is not strong but it exists. I think that I have seen the same also in treemix trees of other papers. Neolithic African admixture seems to be real, but, of course, we can argue that is insignificant, and in any case it is not close to 1/3. However, I think that modern Europeans look less Asian because of Basal.

    In support of the recent two Eurasians theory, this paper includes some interesting alternative treemex trees. In 9.7. A., we see southern populations consisting of Han/Kinh/Dai, Papuans/Australians and mysteriously Nivkh, Ket and Saqqaq; and northern populations consisting of Indians (close to this divide), Central Asians, Europeans and Arctic and Native American populations. In 9.12. A. TreeMix tree, even all Native Americans are in the ”paleo-melanesian” branch. This same pattern is also seen in several other treemix trees.

    ReplyDelete
  54. "In that case, it is absolutely wrong to use Han and MA1 as controls for Native Americans".

    I don't understand or rather don't agree with this. Before Ma1, Native Americans were usually seen as a rather simple subset of East Asians, as they naturally cluster with them in admixture runs and the like, at least at shallow K depths (then they form their own cluster). So using East Asian controls is justified and I dare say quite correct: Native Americans are by ancestry a simple mix of pre-Ma1 and proto-EastAsians.

    In any case that's not the bulk of Raghavan's measures. In the f3 "absolute" Ma1 score exercise, they are compared, like everyone else, with Yoruba and Ma1 only. They still score highest for Ma1 by far, something that seem to be even larger when the IBS data is considered.

    "Native Americans are far more diverged from East Asians than South Central Asians (Pamiri Tajiks, Tajiks, Kalash, Pashtuns, etc) are from the French or any other Europeans"...

    Apples and Oranges. NAs diverged from East Asians maybe 20 Ka ago without further contact except for minor flows across Bering Strait. Additionally they carry heavy (c. 40%) Ma1-like ancestry, which is Western-like. If we take the extreme of West Eurasian internal divergence, which may be the Kalash, they only coalesced as an isolate some 12 Ka ago (allegedly), which is half the time. They also have no extraneous admixture sources.

    Do I really have to explain the obvious? I thought you were oh-so-knwoledgeable...

    "Also, it is very strange to claim that West and South Asians don't have higher ANE ancestry than Europeans based on shared drift".

    After the exchange with Alberto I realize that some Pakistani and nearby populations (but not Iranians nor any West Asians I could discern) may be more in the European range. This is not because their ANE increases, relative to Raghavan's absolute scores, but because European ANE seems to be in need of some haircut, judging of IBS. The amount of change is not large but the rankings are realigned.

    "As noted in the supplementary data of Lazaridis et al., although most Europeans share more drift with MA1 than NE Caucasians, NE Caucasians have higher ANE admixture than any European population".

    I believe that's what David pointed me to earlier in form of unsourced mini-table. If so, the exercise is worthless because we can't expect that the WHG-EEF-ANE triangle works for that area. Per Lazaridis, it does not even work in some European regions (most clearly Sicily and Malta), so much less out of Europe, unless demonstrated otherwise.

    "Please refer to that part of the paper's supplementary data".

    What part exactly?

    ReplyDelete
  55. @Maju

    Yes, if you want to keep the view of ANE = MA1 affinity, then I can only agree with you (my first comment on the thread already stressed this point). And it's probably academically correct.

    But I think that in that case, ANE becomes rather irrelevant, except for Siberians and Native Americans. The rest of Eurasians (i.e, almost all), didn't receive any admixture from that source, for the obvious reasons that we¡re talking about very small populations restricted to a very specific environment and way of life). Kets, Selkups and Mansi are the true descendants of MA1 in Eurasia (though heavily admixed with East Asian populations), but even these populations are bound to be extinct in the next century or two.

    So for practical purposes, we need modern ANE populations in a place where they could mix during the Neolithic with Near Easterners to produce the ANI component ("teal", or whatever we want to call it, but basically a mix of ANE and ENF). Anywhere between Siberia and the near East is possible. I think that around North East Iran is a reasonable option, but you might have a better one (we can call this hypothetical population something else, like Ancient Central Asian or whatever). Then this ANI (teal) population is the one who has contributed a large amount of genes to all South Asian, West Asian, Central Asian populations, and the one who entered the steppe and mixed with EHGs in Yamnaya and Afanasievo cultures (but it came from earlier ones too, we just don't have the genomes yet). And from the steppe to Europe (maybe also from West Asia to SE Europe, but that's open to debate still).

    So whether the ANE in Eurogenes K8 or the "West Asian" in Dodecad K7b (or the "teal" in Haak et al 2015), is what is practical for tracking population movements in West Eurasia. Pure MA1 affinity is just interesting for the history of Siberia and the Americas. That's probably why we've worked around that definition to move onto something more practical.

    ReplyDelete
  56. Correction: by "we need modern ANE populations" I meant "more modern" (than MA1). Like from the Mesolithic.

    ReplyDelete
  57. Can ANYONE tell me about the possible publication time of the Pinhasi study????

    ReplyDelete
  58. @Kristiina: "Maju, several TreeMix trees of this recent paper show an admixture edge from Dinka to Stuttgart".

    Indeed (have been pointing to that for some time myself). IMO that's the famous "Basal Eurasian" thingy but not everybody agrees with me.

    "Neolithic African admixture seems to be real, but, of course, we can argue that is insignificant, and in any case it is not close to 1/3".

    200% in agreement.


    "In support of the recent two Eurasians theory, this paper includes some interesting alternative treemex trees. In 9.7. A., we see southern populations consisting of Han/Kinh/Dai, Papuans/Australians and mysteriously Nivkh, Ket and Saqqaq; and northern populations consisting of Indians (close to this divide), Central Asians, Europeans and Arctic and Native American populations. In 9.12. A. TreeMix tree, even all Native Americans are in the ”paleo-melanesian” branch. This same pattern is also seen in several other treemix trees".

    I don't see how Indians can be "northerner", really. Anyhow this West/East Eurasian divide (with India in the West and Australasia rather in the East) is absolutely normal and the only thing that emphasizes is that, in spite of some SE Asian tour of K2→P1 (and mtDNA N/R), the bulk of West Eurasian autosomal ancestry is "Indian" (although Neolithic and such West Eurasian inflow into India also plays a role).

    Anyhow there are many tree-mix graphs and not sure why would you consider those two particularly reliable.

    I am personally hostile to the two-routes hypothesis. So far I've seen nothing but speculative garbage "backing" it. Haploid genetics and archaeology don't seem to support it in any way. Basically the North Asian route was off reach to Homo sapiens before the Aurignacian (senso lato) and all early migrations went through India (forth and back). Around the early UP, maybe a bit earlier, the two regions (India-West Eurasia vs East Asia), or three regions if including Australasia, finally become differentiated.

    Trees are not ideal to express the complexity of autosomal genetics in any case. They can give some hints though. 9.7 for example is not better than 9.8 (97.99% vs 98.27% of the variance) but the difference is not huge. Yet they produce quite different trees. One of the issues here is that Denisovan input is not considered and that may distort somewhat the position of Australasians (it should pull them artificially towards the root but unsure how it behaves exactly).

    ReplyDelete
  59. @Alberto: I concur that more recent references like Yamnaya or similar are more adequate. They are not "ANE" though but only seem to carry a fraction of it.

    ReplyDelete
  60. Maju,

    Again, I mean no disrespect, but I feel that I'm the one who is explaining the obvious (to you). Much of what your'e saying doesn't make sense, because you seem to be operating under an anachronistic conceptual scheme which has been untouched by the data we've seen from the previous year till now (starting with the Lazardis et al. paper).

    Anyway, Lazardis et al. explicitly contradict your claims in pp. 122-125 of their supplementary data. And the models of Lezgians having higher ANE admixture than any Europeans isn't due to them extending EEF-WHG-ANE to West Asia, because they don't do that! Again, please read pp. 122-125 to get some clarity.

    Regardless, this discussion belongs to early 2014. We've really moved beyond much of this. And even for early 2014, this isn't an informed conversation, but rather one doused in confusion.

    ReplyDelete
  61. Let's say that David's K8 is showing the ANE of the hypothetical MA1-like population that we might find somewhere near Central Asia from around 8-10 Kya. If we call this hypothetical population Ancient Central Asian (ACA, because it's not exactly like MA1-ANE), then we have the following:

    ACA + ENF = ANI (90% teal with a bit of blue_european)
    ANI + EHG = Yamnaya

    Yamnaya works well for Northern Europe and mostly Western Europe (SE Europe might need additional ANI). But for Asia ANI would be definitely better than Yamnaya, IMO. But we still need good ANI samples, first to verify they existed, where they came from exactly, and how exactly did they look like (including their uniparental markers). Hopefully we don't have to wait too long for them.

    ReplyDelete
  62. ”Kets, Selkups and Mansi are the true descendants of MA1 in Eurasia (though heavily admixed with East Asian populations)”

    Alberto, what do you mean with ”heavily admixed”? I have now noticed to my great surprise that Karitiana are 67.4% West Eurasian and Eskimos 66.7% West Eurasian and Kets 66.7% West Eurasian. Let’s see if Kennewick man is the least West Eurasian of all Native Americans if someone would put him through a calculator.

    If ANE=Ma1 and Ma1 had 0% Han component, Native Americans, Eskimos included, are much less East Asian than I thought, but perhaps their mysterious paleo-Melanesian component clusters with ANE and the only East Asian part is the portion connected with yDNA C-M217.

    As you can see, this paleo-Melanesian thing fascinates me, but I fully recognise that it is very far from being proven, and may never be proven.

    ReplyDelete
  63. "We speculate that ANE ancestry in Kets was acquired in the Altai region, where the Bronze Age Okunevo culture was located, with a surprisingly close genetic proximity to Mal'ta."

    It sure is quite old.

    Surprisingly Yamnaya - Afanasevo, Okunevo cultures, were only recent nomads in the region outlined with elevated ANE. More or less we have[in no particular order] Ust’-Ishim -Kostenki 14,- Mal'ta boy. Within the range of Neandertals; extending/overlaping, within the range of Yamnaya and all three samples above, as far as Altai Mountains.

    ReplyDelete
  64. Correction: Kets are more West Eurasian than 66.7% as that is only their ANE percentage and they have also for example East European ancestry inherited from Uralic Maris. Does anyone know how much East Eurasian remains in Kets when ANE is deducted?

    ReplyDelete
  65. @Kristina

    I didn't mean anything that's not seen in any admixture test. Kets can have anything between 40-60% East Asian, depending on how East Asian is defined. Sometimes they use a Siberian cluster, where they get maybe 70%, but that's because Siberians have West Eurasian admixture.

    I would be curious to know more precisely how much East Asian Kets really have. And Native Americans. In theory, it should be possible to make some kind of K2 where one cluster is pure WHG (without SHG) and the other is Dai. That could work, I guess, because ANE should be closer to WHG than to Dai, and any East Asian part should be closer to Dai than to WHG. But I've never seen such experiment.

    ReplyDelete
  66. In K8 they have some 58% East Asian, and in that test the ANE is deducted. It also has a WHG cluster, and a South Eurasian cluster. Plus Bedouin one. So I really doubt that Kets can have much less than that East Asian, but who knows. In any case never less than 40%, I'd say.

    https://docs.google.com/spreadsheets/d/1kd9Q9vFrL1Cra9ayqMYVFKXrUdnThmQJVMtjczLhoTs/edit#gid=74932529

    ReplyDelete
  67. @Alberto

    An f4 ratio using Dai/Onge and WHG/EHG/ANE ought to give the ballpark right answer for East Asian vs "West Eurasian" in Kets (or whoever).

    @Seinundzeit

    "the true source of West Eurasian-affiliated ancestry in indigenous Americans is far closer to the EHG samples from Haak et al., rather than to MA1."

    Why do you say that?

    MA1, EHG; Karitiana, Chimp -0.0002, -0.033

    Looking at a set of D stats for over 200 modern and ancient samples, I see that precisely one of them is significantly closer to MA-1 than to EHG, and that is of course AG2.

    Only a couple of modern humans are positive for (MA-1, EHG; X, Chimp) at Z < 1: Anzick-1 and Australian Aborigines. At Z > -1 we have for modern populations only Papuan, Bougainville, Athabascan, Bolivian, Maya, Mixtec, and Surui.

    Almost everyone is at Z < -1, including Ju Hoan, Biaka, Bantu, Han, and Dai. This includes some Amerindians groups, but they are generally less negative than Sub-Saharan Africans. Or macaque monkeys, for that matter.

    So I really don't see that the ANE ancestry in Amerindians is much closer to EHG than MA-1. More likely the other way around.

    ReplyDelete
  68. @ capra,

    In Haak et al., the qpGraph modelling has Native Americans receive their West Eurasian admixture from a North Eurasian population that constitutes a clade with EHG in comparison to MA1 (if I'm recalling things correctly). I'd say that the qpGraph modelling is determinative in relation to anything we see with d-stats, since it is the meat and potatoes of the Lazaridis et al. paper.

    To test this further, it would be very interesting to see qpAdm output with the Karitiana as a mix of Dai + EHG and Dai + MA1. We could then easily compare and evalute both models in terms of stats.

    ReplyDelete

  69. "An f4 ratio using Dai/Onge and WHG/EHG/ANE ought to give the ballpark right answer for East Asian vs "West Eurasian" in Kets (or whoever)."

    Can anyone calculate that? I definitely cannot.:-)

    Does ANE contain any East Asian in the above calculation? I remember that before in many discussions it was often emphasized that Ma1 did not have any East Asian.

    "the true source of West Eurasian-affiliated ancestry in indigenous Americans is far closer to the EHG samples from Haak et al., rather than to MA1."

    I agree that common sense is against that. Native Americans should have reached Beringia before 20 kya, so their arrival is close to Ma1's age. Samara samples are only 5 (?) kya old and contain a lot of recent drift that cannot be shared with Native Americans. Ma1's geographic location is also much closer to Beringia compared to Samara.

    ReplyDelete
  70. EHG and MA1 are equally close to Native Americans.

    ReplyDelete
  71. When I claim that the ancestors of Native Americans should have been in Beringia before 20 kya, I refer to this upcoming paper:

    http://meeting.physanth.org/program/2015/session13/lee-2015-ancient-dna-analysis-of-human-remains-from-the-siberian-arctic.html

    The abstract says that ”Archaeological evidence illustrates the presence of human occupation in the Arctic Siberia by 27,000 years before present (YBP)” … ”We extracted DNA from the skeletal remains ranging from around 27,000 YBP to as recent as the 18th century AD…”. ”We successfully identified maternal lineages for five individuals that include haplogroups A2 and C4.”

    ReplyDelete
  72. Wow good effort, despite it being only mtDNA

    ReplyDelete
  73. There is a good chapter posted on academia about the Americas
    It discuses the wide ranging estimates for colinization- from 25 to 13 kya, the role of climate and sea levels etc .
    https://www.academia.edu/14928240/The_Initial_Colonization_of_North_America_Sea-Level_Change_Shoreline_Movement_and_Great_Migrations

    A later date might explain why amerindians don't form a Clade with Mal'ta.

    ReplyDelete
  74. Sf11 from Mesolithic Sweden had pre-I1. Geneticker also did analysis of Pitted Ware and a single Funnel Beaker Swede. They all had I. I added them to my Ancient Y DNA spreadsheet.

    https://docs.google.com/spreadsheets/d/12G2cfjG0wHWarsl5bB99ridFmvUWzqlZfZ6_e_R6oIA/edit#gid=0

    I1 was also found in Neolithic Hungary at around the same age. Sf11 was pre-I1 so an uncle to modern I1 not an ancestor and I would guess the same is true for the I1 from Neolithic Hungary.

    It's obvious many differnt I1 lineages existed in Mesolithic/Neolithic North Europe, and a *single* I1 lineage went through a founder effect and became especially popular in Late Neolithic Scandinavia(Proto-Proto-etc. Germans?) by 2000 BC.

    ReplyDelete
  75. @Kristiina

    I hope they got some autosomal DNA, preservation ought to be pretty good there you'd think.

    I don't have the tools to calculate it, but in theory I think f4(Ket, Dai; Loschbour, Mbuti)/f4(MA-1, Dai; Loschbour, Mbuti) would tell you how much ANE is in Ket, except that probably everyone is too admixed really. Could try it with Onge, K14, and other groups as well to see how robust the result is.

    @Seinundzeit

    AFAIK ADMIXTUREGRAPH tests models with f4 statistics. There were a bunch of possible models in the paper, with all possible relationships between MA-1, Karitiana, and Karelia_HG, including a trifurcation, because MA-1 and Karelia_HG are equidistant from Karitiana in f4 stats.

    They were running qpAdmix with Karitiana as an outgroup, they didn't use it for Native Americans that I can find.

    My point was that since EHG shows up as closer than MA-1 to almost every modern population in D stats, including San, Pygmies, and even frigging monkeys, that there might be some kind of artifact diminishing MA-1 affinity. But even if not, EHG isn't closer to unadmixed Native Americans.

    ReplyDelete
  76. Yes, I hope that too. Ma1's population's possible admixture with a Beringian population should emerge from a comparison of Ma1 and that 27 kya sample.

    This week has been really eyes-opening. So, West Eurasians seem to have emerged from their admixture with Near Eastern populations and ancient North Eurasians cluster with East Asians or at the root of East-West divide (e.g. Ust Ishim, Ma1). Quite logical if we consider the origin of yDNA K! That also explains why it seems to be so difficult to distinguish ENA from ANE. However, basically ANE should derive from South Asia and East Asian from East Asia. When we get paleo data from East Asia, it will be easier to make that distinction.

    ReplyDelete
  77. I found out recently that a few samples from this site in Bulgaria were tested, but they didn't make it into the Haak paper.

    http://www.archaeology.wiki/blog/2013/01/03/ancient-civilization-uncovered-in-bulgaria/

    ReplyDelete
    Replies
    1. why ? Poor quality ?

      Btw That brief article is problematic, as is the entire commonly held narrative of Cooper age decline in SEE

      If the civilization was "destroyed" by invaders c. 4100 BC, then why don't the earliest kurhans appear until 3000 BC ?

      If someone is going to suggest that they went back to "HQ" to regroup and resupply (for 1000 years), don't bother :)

      Delete
  78. Apparently the samples weren't ready in time. I suppose they had to choose a moment when to submit the preprint, and whatever was ready by that time made it into print.

    I don't know much about that link. It was the first relevant thing I saw in Google. I'll withhold my judgment as to its accuracy until I see the aDNA results from Neolithic to Bronze Age Balkans.

    ReplyDelete
    Replies
    1. Oh I'm not saying that there wasn't a movement from the steppe to northeast balkans, but according to these most recent carbon datings, the scenario is different to that usually painted. Maybe the "Yamnayans" were peaceful herders filling long vacated land. (?)

      Delete
  79. David,Alberto,Mike etc
    Who is Johannes Meller? Is She Related to genetics?.

    ReplyDelete
  80. Where did you get that name from?

    Don't you mean Johannes Krause? And that's a he, by the way.

    http://www.shh.mpg.de/28671/research_outline

    ReplyDelete
  81. Thanks David,
    Mike, Do you think the R1b dominance of Yamnaya so far reflects a founder effect, which replaced earlier groups?

    ReplyDelete
    Replies
    1. It's complex, and you need to define what you mean by "R1b"

      But basically, i think r1b is overall from the south, but M269 might have risen on the steppe itself. Certainly, M73 looked "at home" on the steppe

      Delete
  82. David, No i don't mean Krause, Its Meller....

    ReplyDelete
  83. I don't know who Johnnes Meller is.

    ReplyDelete
  84. Then I wait for others, if they can give a detailed info.

    ReplyDelete
  85. The only "Mellars" I know is Paul- the eminent Paleo-archaeologist

    ReplyDelete
  86. @Sein: Do you mean: "Table S14.15: Admixture proportions for Near Eastern populations fit as a mixture of the ancient East and Ancient North Eurasians. A lower bound that can be obtained via the ratio f4(Test, Stuttgart; Karitiana, Onge) / f4(MA1, Stuttgart; Karitiana, Onge) is also indicated"? (I'm using the pre-pub, so it is in a different page: 123).

    I don't think that is acceptable as test of anything: the reference for Near East is not clearly described, although I'd expect something like Naqab Bedouins or an extrapolation based on them is what they consider "Near Eastern". That's an assumption that I just do not accept as valid for all Ancient West Asia. In my understanding the "Gedrosian" component was already there, although maybe with different, more geographically restricted distribution than today.

    Also nothing Indian (Ancient East Asian is from context derived from Onge or Dai, which are SE Asian populations).

    For me all that is just rubbish. At best a playful exercise within their assumptions without serious meaning, never something that can be taken seriously.

    "Regardless, this discussion belongs to early 2014. We've really moved beyond much of this. And even for early 2014, this isn't an informed conversation, but rather one doused in confusion".

    In my understanding you guys have been discussing in the wrong parameters and hence you're reaching to wrong conclusions. Direct "absolute" measures of ANE clash with all those ideas, in Raghavan as in Ayub & Mezzavilla. Just because you guys are entrenched in some wrong ideas via "group thinking" in whatever forums you do that vicious circle practices doesn't mean that you are right, only that you slap each other on the back too often.

    So less group-bullying of people with different ideas (I got vaccinated against that in school, it won't work with me) and more critical thinking and comprehensive discussion, please. What I say, and that is the challenge I issue, is that your camarilla (whoever they are) have been thinking wrong on this matter for two years, judging on what you say, producing wrong results with wrong methods and in general diverting scientific advance to a Piltdow Man of sorts.

    Of course, I can also be wrong in some aspects. In this sense I want to thank Alberto again for the feedback because it does suggest that some corrections are needed to equilibrate "absolute" ANE measures for Europeans (a bit too high) vs West/South Asians, although it does not alter the overall results dramatically nor produces high ANE anywhere in West Asia.

    ReplyDelete
  87. @Alberto: "But for Asia ANI would be definitely better than Yamnaya, IMO."

    But ANI is Neolithic flow from the Zagros area most likely. It's like the EEF/Sardinian component in Europe, just that it has suffered much less apparently. ANI is roughly the same as Gedrosian, Baloch-Caucasian (specifically the Baloch part) or Highland West Asia: it has nothing to do with Siberia or the steppes (except that some of it seems to have migrated north and been incorporated into Yamna).

    Anyways, Sintashta should be directly ancestral to the Indo-Iranian peoples in terms cultural and archaeological (not necessarily genetic, of course). How does Sintashta perform? I bet that not much different than Yamna.

    Something that we can notice in the limited West Asian aDNA samples we have, is that Mittani people brought some Indian mtDNA with them, hence they were expanding from (near-)Pakistan rather than from Central Asia.

    Of course I still support a conquest of late IVC from Central Asia c. 1900 BCE (Cemetery H) but afterwards the genetic pool of Indo-Aryans seems to have been strongly "Indianized", as corresponds to a conqueror minority.

    ReplyDelete
  88. Maju, I'm sorry but what you said in respect of IIR is a complete hogwash, lord save the poor old fella....

    ReplyDelete
  89. Indo-Iranians, BTW why are so reluctant to insult the Indian culture, tell me do you have any personal agenda? I mean you surely don't believe in religion or fate but i think you will have very sad ending.

    ReplyDelete
  90. I don't understand your POV, Nirjhar. Sounds Hindutva to me, i.e. Brahmin oligarchic political-religious ideology, a fundamentalism of sorts. I don't feel I've disrespected anyone but I don't care if reactionary fanatics feel offended.

    I don't have any personal agenda other than global communist revolution (but that is well known and has nothing or very little at best with my fascination with prehistory and anthropology).

    As for my fate it is the same as everyone: to die. I hope it is a bit dignified but in any case the individual end is something we all must face and that never feels "happy".

    ReplyDelete
  91. So you are a communist?, I am a Rationalist not Nationalist.
    Your Preconceptions make me vomit, anyway i was a Kurganist a while back, I found its ridiculous and moved on......
    Of course you will feel happy if you get rid of your dogmas, you are an intelligent person....

    ReplyDelete
  92. "I am a Rationalist not Nationalist".

    Then why do you claim I offended Indo-Iranian peoples as a whole? It makes no sense!

    (Also I'm communist because I'm rationalist but particularly because I'm humanist - off topic).

    ReplyDelete
  93. You are Insulting them and with the Indian Culture also, you are trying to make them what they aren't, Its just utterly disgusting.
    Humanism and Communism are far off things...

    ReplyDelete
  94. 'Uralic languages formed during the Bronze Age/Iron Age upheavals from a language that was close to Proto-IE and has several paleo-Siberian languages as substrate languages.'
    The study implies that ASTt was paleo-Eskimo and an important cultural/genetic nodal point to both pan-american integration and technological advance (bow, pottery), that thus affected all the Americas. It may be deduced that paleo-Eskimo represented the Siberian component already in MA-1 but different from Amerindian and East Asian. The origin of paleo-Siberian was presented along the lower Lena (Bel'kachi culture, 3200-2100 BC), but I don't think there is any reason to assume it's expansion preceded their newly achieved technological edge. Proto-Uralic had words for bow and arrow, though to ascribe their advance to a paleo-Siberian substrate appears quite anachronistic to me.

    ReplyDelete
  95. @Nirjhar: "You are Insulting them and with the Indian Culture also, you are trying to make them what they aren't, Its just utterly disgusting".

    I am not. You have not provided any instance of any such alleged insult. You're actually insulting me with your unjustified claims.

    ReplyDelete
  96. You are blind with your dogma, i honor your knowledge in Genetics but obviously you have discriminated the Aryans, Just like Nazis did to Jews...

    ReplyDelete
  97. 'I don't have any personal agenda other than global communist revolution (but that is well known and has nothing or very little at best with my fascination with prehistory and anthropology).'
    Hmm, nevertheless politic c.q. anti-western, hence worth fighting for. Kurganists have a political agenda, also very well known. Why some people still think a Kurganic world will bring back equalty to this world? And does it? Fascinating.

    ReplyDelete
  98. @Rokus: I'm Western European, not at all "anti-Western". Just different kind of Western ideology. Marx, Durruti, Lafargue and Likiniano were all Western. Or are you telling me that Humanism is "anti-Western"? Please!

    @Nirjhar: you are still not explaining what did I say that could be offensive. Your accusations are ridiculous and you are very much insulting me without any basis and I would punch your face if this conversation was face to face, believe me.

    ReplyDelete
  99. @Rokus: "Kurganists have a political agenda, also very well known".

    Not really. The Kurgan model is the only model that coherently explains the expansion of Indoeuropean languages. It is as purely scientific as it can get. That does not mean that maybe some people may use it to further an agenda but I assure you it's not my case. For me understanding IE expansion (and the Kurgan model is the only one that makes sense for this) is about understanding recent paleohistory (culminating in historical periods, even recent ones such as European colonialist expansion) about understanding not just the place of Indoeuropean languages and associated ethnicities but also that of pre-IE languages and associated ethnicities like Basque, Caucasian(s) or Dravidian. In brief: about understanding our shared prehistory, roots, etc.

    ReplyDelete
  100. Maju,
    Well I will not do that to you because i respect you, you are old, but of course which way you are implying on Indian and Aryan History, I am extremely disappointed,In our culture patience is very important with constant effort...
    Communism is not Humanism, Its a failed concept, but Humanism is natural and all prevailing.

    ReplyDelete
  101. Maju,
    ''The Kurgan model is the only model that coherently explains the expansion of Indoeuropean languages''
    To you perhaps but surely not for those who has scientific attitude and studied the basis of its conclusions, especially the case of Asia.

    ReplyDelete
  102. I don't care if you're "disappointed" or not, Nirjhar. I care that you just insulted me calling me "nazi", something that for me is like the worst offense and of course has nothing to do with my stand, which is a polar opposite.

    Also you have failed to clarify what is that I said that you feel "insulting" towards Indo-Iranians. It all seems totally weird and absurd and definitely obscure.

    And the Kurgan model also works for Asia, just that not in the demic-genetic sense that some people would expect or like. Indoeuropean languages and the corresponding ethnic identities are cultural elements not "blood" (genes). Many peoples around the world have seen their identities quite radically altered through the ages, it may be sad or whatever but I assure you it is not my fault.

    ReplyDelete
  103. Maju,
    ''Also you have failed to clarify what is that I said that you feel "insulting" towards Indo-Iranians.''
    well you will never see it, because you don't have a free intellectual mind, if you find that offensive then i apologize but that's the case.
    You have zero idea Maju Zero idea, your conceptions are narrow and circular, they lack broadness and empathy, which can only be learnt by unbiased mind, hence your ideas are not much different than of the nazis...

    ReplyDelete
  104. I'm going to ignore you hereafter, Nirjhar. Your insults and outlandish unspecific claims are just the straw.

    ReplyDelete
  105. Maju,
    I feel terribly sorry for you.

    ReplyDelete
  106. 'The Kurgan model is the only model that coherently explains the expansion of Indoeuropean languages. It is as purely scientific as it can get.'
    Come on, you are intelligent - even being as wrong as you could be.
    A theory or hypothesis is only "scientific" if it meets certain criterions, one of those being Karl Popper's falsification criteria. Hypotheses that can't be rejected based on a prediction don't meet this criterion and should be dismissed as unscientific and pseudoscience. The number of violations of Kurganism to this principle in modern publications is countless. The list of falsified predictions includes the steppe origin of blue eyes (already present in Western Europe long before), lactase persistence (that apparently missed PIE expansion), white skin (already existed in LBK hybrids), YDNA R1a (null in Yamnaya, what a joke!!!), mtDNA H (already found in Mesolithic Iberia), domesticated horses (actually derived from a western European variety), absence of IE languages in Bronze Age western Europe (highly hypothetic and contradicted by valid evidence), etc. etc. Of course none of these findings escaped stubborn denials, since Kurganism already has another reality that "coherently explains the expansion of Indoeuropean languages".
    By now the predictions of a Steppe origin are violated by autosomal DNA that apparently favor a more eastern origin than originally predicted, with constructs like 'eastern Neolithic' that may divert the attention from 22% ANE already in Motala12. Why, did anybody ever truly expect that Stuttgart EEF (up to 80% WHG) was ancestral to all European EEF, rather than being a local hybrid? The recent find of Greek Neolithic genomes confirmed that some of LBK's most celebrated 'innovations' like a white skin were actually locally acquired in Central Europe. Moreover, this implies that the eastern Neolithic WHG component was independenty acquired from Stuttgart in Yamnaya and Corded Ware, so where did this come from if it wasn't EHG? A Yamnaya origin is contentious at the very least. Many Kurganists prefer to have the discussion closed and build on their house of cards, but unfortunately it isn't.
    But why so few genetic discoveries ever confirmed Kurganist predictions, as far Kurganism ever bothered to predict anything? Pseudoscience is typical for closed systems that may be religious, political or overly comprehensive such as Kurganism and other isms or Chomsky. Incredible this way of thinking still exists, and pseudoscience thrive more than ever. Is this the end of civilization, or does at least your version of Kurganism have a falsification criterion all of its own that I am not aware of?

    'about understanding our shared prehistory, roots, etc.'
    I would be happy if Nordwestblock and Swifterbant could ever share the same level of understanding.

    'And the Kurgan model also works for Asia'
    Not if this involves a Yamnaya origin.

    ReplyDelete
  107. Not talking about Maju, But I'd imagine that Basqo-centric occidental autochthonists need the kurgan theory. To them , "the invading kurgan hordes" provides a nice antithesis to the palaeolithic- basque-R1b mythical people they created.

    ReplyDelete
  108. @Mike Thomas

    "That brief article is problematic, as is the entire commonly held narrative of Cooper age decline in SEE

    If the civilization was "destroyed" by invaders c. 4100 BC, then why don't the earliest kurhans appear until 3000 BC ?"


    One possibility is three adjacent regions: A, B and C and pastoralists in region A raid B causing population B to move away to region C.

    End result: depopulated region B used as summer pasture by the pastoralists while they slowly repopulate it and civilization C collapsing under the weight of refugees from B.

    ReplyDelete
    Replies
    1. Well is possible in the world of Grey, where evidence is not solicited, but rather one drums up fantastic scenarios inspired by John Wayne movies :)

      Delete
  109. @Rokus: Popper did not invent science, he just died some years ago what makes him a baby of science rather than the father of science!

    Anyhow: those "predictions" are not in the Kurgan model, which first of all (read Gimbutas, please) is about elite domination, rather than simple demographic expansion. What you're talking is about certain demicist extreme interpretations of it, of which I declare myself and Gimbutas innocent of.

    What you are talking about and what I'm talking about when we talk of Kurgan Model is quite different ideas. I'm quite strict to the original thing, you are surfing on the waver of some demicist reinterpretations.

    ReplyDelete
  110. @Mike: actually I've got criticisms for paying "too much" attention to Indoeuropeans by some more "extremist" vasco-centrists. Also for claiming that the Basque ethnogenesis is Neolithic, which seems to be a bit "disappointing" to those. Many don't even like the relationship with "Mediterranean" peoples like Iberians or Sardinians, go figure! The Basque ethno-ideological paradigm (largely a product of bad "autistic" linguistics) is largely entrenched in uniqueness, anything non-Basque startles them as disturbingly "foreign". Luckily most Basques are more open-minded than that but there is still something of all that, yes.

    Anyhow, the R1b theorization was done largely by English (Sykes, Oppenheimer) rather than Basques. Let's not blame Basques for something that has been re-constructed in other places mostly.

    ReplyDelete
    Replies
    1. @ Maju

      Fair enough. My point was perhaps best directed at Gimbutas herself. She created the "Kurgan people" - patriarchical, Warriors, aggressive, nomads as the exact antithesis of her "old europe" farmers. Her theories derive more from her feminist, socialists ideology than fact.
      To people like me, who are intelligent and unbiased in regard to what prehistory they wish to envision for their "IE ancestors", her reconstruction was half-true at best.

      I've a feeling that more evidence will further show that the kurgan theory will require major
      modification, if not abandonment

      Delete
  111. @Maju, you didn't get the message. A model without falsification criterion is pseudoscience. There is no excuse for Gimbutas' followers since Einstein already applied the principle with sound predictions that supported his results. Popper was a philosopher that merely described good practice, clearly waisted on the current lot of crazy Kurganists.
    'I'm quite strict to the original thing'
    The failed Kurganist reinterpretations were meant to save or confirm the model, because the folly of preaching conservative Kurganism in the modern world was repeatedly recognized. You want to stick to 'Elite dominance' as your/Gimbutas criterion of Yamnaya expansions, even now this can easily be rejected (falsified) based on YDNA?
    At least the male elite of Yamnaya society were the exclusively carriers of R1b(xL51), what rules them out as the elite that also dominated roughly related cultures credited for being ancestral to probably all IE cultures except Armenians. To assume that R1a and R1b-L51 should have originated from the lower echelons of Yamnaya society violate both the available evidence and Gimbutas' 'original thing'.

    ReplyDelete
  112. @Mike Thomas

    Well it's testable - if correct there should be a pattern of evidence of depopulation on the border of pastoralist expansion and conflict / overpopulation / collapse on the *opposite* borders of pastoralist expansion.

    So say there were five regions A to E with the pastoralists in A then if correct the pattern of evidence would go in a sequence

    B depopulates
    C collapses
    A expand into B
    C depopulates
    D collapses
    AB expand into C
    D depopulates
    E collapses
    ABC expands into D
    etc

    ReplyDelete
  113. @Mike: Gimbutas and some followers were no doubt carried away by some polar exaggeration. However, if you take that part of anthropological interpretation with the necessary pinch of salt, you find that it still seems to be reasonably real. Indoeuropeans were indeed annoyingly patriarchal in the Mediterranean context, while other peoples like Etruscans, etc. were quite apparently much more gender-egalitarian (although still somewhat patriarchal, soft-patriarchal). So that partiarchal extremism of women-at-home that culminated in Islamic clothing styles (which were borrowed from Byzantine Greece) is indeed something that IEs brought with them from the Northeast (although there are other sources like Semitic tribes, etc. - just read the Bible to get an idea). However it's possible that it is not merely a part of their steppe pastoralist cultural baggage but something reinforced precisely through conquest (hence some IE peoples which persisted in harsher, wilder, areas were maybe a bit less extremist on this).

    Mythology is also interesting: the pre-IE divine couple (rather than just mother goddess) religion is everywhere: Basque mythology, Greek mythology, Shaktism in India, Taoism even in a more abstract or philosophical form. It's truly impressive how universal this early dual-monotheism is. In contrast to this "Neolithic religion", whose values are fertility and community, the IE religion is one of winners, even if cheaters or cruel or unjust. It is a materialistic religion of success, military success in particular, raw power.

    It's not as simple as mere patriarchy on/off. Patriarchy or something very similar surely existed earlier but it was more balanced and community oriented and women had important roles not just as mothers and wives.

    "I've a feeling, contrary to what many people still feel, evidence will soon show that the kurgan theory is completely false".

    Science is not made out of feelings. That's Wonderland or, in plain terms: wishful-thinking. The main problem is that archaeology is very strongly supportive of the Kurgan model and no alternative model makes good sense. Additionally it fits well with whatever linguistic science can reconstruct and now there is also some significant archaeo-genetic evidence.

    Rather than demonstrate the Kurgan model as false (which seems very hard) the only possibility lays on forging a new credible and internally consistent theory that happens to be even better. I don't see anything like that happening because there's nothing to support any such alternative theory. The best attempt probably was Renfrew's Neolithic hypothesis but this one has been proven wrong on several grounds already (and was never nearly as good as his colleague's).

    ReplyDelete
  114. @Rokus: https://en.wikipedia.org/wiki/Scientific_method

    Not every scientific field can make clear-cut "testable predictions", that's something that mostly applies to Physics and other exploring disciplines of the fundamentals of the universe. It could apply to Economics but this "science" is so extremely ideologically polluted that in practice it does not, else 99% of economists should be considered non-credible because they almost systematically fail their predictions, which are largely based on the wishes of whoever pays them.

    What materially-based Prehistory can say is this is what we can gather based on the existent evidence, this seems most plausible and Occam's Razor favors it strongly. As research advances further evidence can force the model to be abandoned or merely reformed or even hard-confirmed. As the Kurgan model (original Gimbutas version, roughly the one I support) does not claim a strong demic expansion, your argument is pointless: it only demonstrates false those interpretations of the Kurgan model that over-emphasize the demographic expansion aspect but not the general model which is only about ethnic identity, language and other aspects of culture such as religion, etc.

    Instead the model of Basque Paleolithic Continuity has been effectively debunked by the recent archaeo-genetic data, suggesting that a Neolithic origin model is the most correct one, even if some details are still obscure. That's because this model relied heavily on genetics (also partly on archaeology and that has also changed somewhat recently but mostly on genetic reconstruction á la Sykes & Oppenheimer). But the Kurgan model does not and, in fact, has posited since its origins a model of conquest and assimilation rather than (mostly) settlement.

    So you are not discussing the Kurgan model as such but a particular secondary re-interpretation of this one.

    Anyhow, String Theory, for example, is not testable, it has been re-created once and again on the scraps of previous models and still enjoys widespread acceptance in the scientific community. There are Popperian fundamentalists that argue like you do against it but others don't seem to care. If you can falsify X variant of String Theory, the Y variant can still be the correct one, and if this one is demonstrated false, maybe there is a Z variant that still holds up. In brief: even if the basics are the same, these are different theories for all practical purposes and each one must be judged on its own merits.

    Similarly the demic-replacement version of the Kurgan theory is a different one of the non-demic version, which is much closer to the original one anyhow, as the emphasis was always on conquest, subjugation and eventual assimilation of other peoples.

    ReplyDelete
  115. Bullshit.

    The Kurgan expansion was a massive demic expansion, which saw several population turnovers in some regions, often of Kurgan groups themselves.

    And just because not all of the Kurgan-derived groups built Kurgans, doesn't mean they weren't part of the expansion.

    ReplyDelete
  116. 'Rather than demonstrate the Kurgan model as false (which seems very hard)'
    The Kurgan model does not offer falsification criterions, and this very impossibility to disprove the model with evidence defines it as pseudoscientific. Just like religion: Does evolution disprove the existence of God? The existence of extremely contradictory rituals and believe systems? That Earth is a round planet? No, because religion is not scientific. But you present this failure as an asset. You are caught by the inherent circularity of unscientific thinking.

    'However it's possible that [being patriarchal] it is not merely a part of their steppe pastoralist cultural baggage but something reinforced precisely through conquest (hence some IE peoples which persisted in harsher, wilder, areas were maybe a bit less extremist on this).'
    Sure, war and upheaval change a culture. Hence, it should be expected that IE regions that never participated in the steppe conquests, remained less so. For instance, Germanic mythology has a violent upper layer induced by their Roman wars and a deeper animist layer that can be traced also in Greek and Indo-Iranian mythology. The latter was completely ignored in the Kurgan tradition. For instance, Odin transformed from creator and poet into a warmonger only during the Migration Period. Hence, Gimbutas' ideas of a patriarchal IE society may have been nothing but a subset of PIE-derived cultures rather than 'the original thing' at all. Gimbutas introduced a dualist world where 'Bronze Age' IE cultures and 'Neolithic' agrarian cultures could not be reconciled anymore within a single origin. Now even the Steppe cultures attest a considerable proportion of Neolithic DNA it should be clear this is just another Kurganist folly.
    What you mean with 'harsher, wilder areas' that were less extremist? Normally it is just the other way around. Hence, less extremist IE should be expected in some regions included in what Gimbutas pushed aside as 'Old European'. An example may be the nordwestblock Triple Godess, by its nature and context completely compatible to IE mythology and still related with fertility and the cycle of life and nothing to do with patriarchal ideas.

    ReplyDelete
  117. Indeed, the String Theory does not predict anything, that is why it is so utterly useless. It only confirms what we already know, nice. However, even in physics there is still more we don't know and String Theory will never help to find out what.
    But dear Maju, contrary to the String Theory the Kurgan Model doesn't even confirm anything! Every new find on this field, including genetics, continues to be utterly surprising and contrary to all previous expectations. However, since the Kurgan Model is pseudoscience, it can always be reinterpreted afterwards, and that is what happens all the time. An asset of a 'solid system', or rather the folly of a crazy bunch of idiots?

    ReplyDelete
  118. @Rokus

    "A model without falsification criterion is pseudoscience."

    In an ideal world... In the real world of "science," falsification criteria are rarely advanced and when they are, a theory gets quickly falsified but still remains a flagship scientific theory (now 90% correct). So it's not an either-or situation but rather a continuum of more or less falsified or supported theories.

    "The list of falsified predictions includes the steppe origin of blue eyes (already present in Western Europe long before), lactase persistence (that apparently missed PIE expansion), white skin (already existed in LBK hybrids), YDNA R1a (null in Yamnaya, what a joke!!!), mtDNA H (already found in Mesolithic Iberia), domesticated horses (actually derived from a western European variety), absence of IE languages in Bronze Age western Europe (highly hypothetic and contradicted by valid evidence), etc. etc. Of course none of these findings escaped stubborn denials, since Kurganism already has another reality that "coherently explains the expansion of Indoeuropean languages"."

    Nicely summarized! With the Anatolian theory being an even worse alternative, what's left?

    ReplyDelete
  119. German, I like your sense of humour!

    ReplyDelete
  120. @Kristiina

    Thanks. I like how you spell "humor." :)

    ReplyDelete
  121. EU is under the British rule. :-)

    ReplyDelete
  122. God forbid we start spelling things like the US

    ReplyDelete
  123. @Mike Thomas

    "Mike" is as US as it gets. Everybody's Mike or Bob around here. I thought Brits tend to keep their first names unsullied by users' tongues. Kristiina, Rokus - those are good European names. German, too :)

    But enough of that... Mike, which theory of IE dispersals do you favor in the light of all the recent whole genome and ancient DNA data?

    ReplyDelete
  124. German im flattered you ask
    Mine is a disappointingly boring "agnostic" (Id guess one would call it) one.

    I appreciate that there was a lot of movement and admixture going on in the LN-EBA, and the steppe played a major role..But I think we've only uncovered part of the story, so far. In the end, it's going to be more complex than seen now, which will inevitably lead to a dead end as far a Y haplogroup chasing is concerned.

    So ultimately, it will fall back a need to develop a theorem which incorporates sociolinguistics, psycholinguistics and contact linguistics into the traditional comparative - phylogenetic IE framework, and uses the archaeological evidence as analysed by bronze age specialists rather than by indo-Europeanists.
    Ie a wholly new Model.

    ReplyDelete
  125. @Mike Thomas

    This is a good framework to have in order to filter through the ever changing datastream. IE language material will be better off, too, if analyzed by specialists other than Indo-European linguists. :)

    ReplyDelete
  126. @Rokus:

    1. You can see in Davidski's reaction how his version of the Kurgan theory and my own are completely different. He's dead set to defend the demic replacement version even without support because he is not satisfied with being the adopted slave of the ancient masters, he wants to be the ancient masters in person, their direct descendant to be precise. Megalomania, I call it.

    2. The falsifying criterion is, I believe, up to the challenger to decide. If you can demonstrate that X assumption of the original theory is wrong, then the theory will have to be revised or even dumped. But the theory has many aspects, you are not going to debunk it without studying the subject. Just because you don't like that ugly building, it's not going to fall down, you know.

    It has nothing to do with religion or other superstitions. This is as good as science does and it's up to you to prove it wrong, in full or part.

    "Germanic mythology has a violent upper layer induced by their Roman wars"...

    I don't think it can be claimed that it was induced by the Romans in any way. It's old, intrinsic and very much rooted in the shared Indoeuropean legacy. I think you're whitewashing Germanic mythology.

    "... and a deeper animist layer that can be traced also in Greek and Indo-Iranian mythology".

    Romans also had a vast animistic imaginary world. It's very common but it's not the bulk of the religious corpus. In Germanic mythology for example we can see also the Old Neolithic Religion as "enemies", notably Hel (Gaia, the Neolithic goddess) and Jormungandir (Sugaar, the snake or dragon, the male aspect of that ancient monotheism). That's the core of the defeated and subjugated mythology and is present everywhere in Europe. But only among Basques it remained as main faith until christianization.

    "For instance, Odin transformed from creator and poet into a warmonger only during the Migration Period".

    Then how do you explain that the same form exists in Scandinavia, which was not affected by that phase? I don't say that gods don't change, they do, but I think you're overinterpreting here.

    "Hence, Gimbutas' ideas of a patriarchal IE society may have been nothing but a subset of PIE-derived cultures rather than 'the original thing' at all".

    It's not just Gimbutas, there are other theorists. I for example drink a lot from J.I. Hartsuaga ("Compared Basque Mythology", not available in English AFAIK), who argued very solidly precisely for the "winner" core of the IE mythology. He may have exaggerated somewhat but I'm pretty sure that he was also onto something.

    Anyhow, how do you explain that precisely in the Germanic pantheon we find so very few goddesses, with so little power, and that Valhalla was denied to women (and everyone who was not a warrior). There are many more goddesses in Greek mythology even if ancient Greeks were surely as extremist patriarchal, probably more, than ancient Germanics.

    In any event the part (i.e. Germanic mythology) is not the whole (IE religion in general terms). Each subculture evolved separately and through its own distinctive paths.

    ...

    ReplyDelete
  127. ...

    "Gimbutas introduced a dualist world where 'Bronze Age' IE cultures and 'Neolithic' agrarian cultures could not be reconciled anymore within a single origin".

    Well, there was still Neolithic continuity in the Bronze Age this side (west) of the Rhine, mind you. However some of the same developments that happened towards the East also happened here but on its own terms. I'm talking of the evolution of oligarchies, division of labor and exploitation, what is definitely not something exclusive of IEs. That happened everywhere since the Chalcolithic, however, while East of the Rhine, the core of this transformation became the IE conquest and its legacy, West of the great river it was done on different local roots. Similarly, Etruscans, which are rooted probably in a third different component from the Aegean (Pelasgian?), also developed slavery, elites, warriors and what-not... but their women retained a more prominent and dignified status, which made Romans consider them "immoral".

    In any case it is more and more, and not less at all, apparent that there were major changes with the arrival of Indoeuropeans to "peninsular Europe" (i.e. West of the Odessa-Gdansk isthmus). So I don't know why are you accusing Gimbutas when the newest archaeo-genetic data from Central and Eastern Europe supports her at least in most aspects: there was some sort of major change, even to some extent demographic, in the Chalcolithic precisely with the arrival of Kurgan peoples from the East. There is discontinuity and usurpation, even if we may not like it.

    "Now even the Steppe cultures attest a considerable proportion of Neolithic DNA"...

    Nothing to do with "peninsular European" Neolithic. Its some other element contributing to their genesis from West Asia but clearly distinct from what affected the bulk of Europe. This only helps to clarify the origins of the earliest Kurgan cultures (Samara, Khvalynsk), which were before not really understood.

    "What you mean with 'harsher, wilder areas' that were less extremist? Normally it is just the other way around".

    IEs established themselves earlier in Central and North Europe and also in the North or NE parts of the Balcans before they did in the Mediterranean proper. So the conquest of Greece, Italy, etc. was a secondary process, needing of the brutal subjugation of new conquered majorities and particularly their women. I'm thinking in terms of what Castilians did in America for example (before the "whitening" of the 19th century), establishing a true race-caste system in which the closest ones to European blood purity were always higher in status. Women were almost invariably local, first native, later native and admixed and were often treated as chattel, as not true companions but flesh utilities. I believe that much of the same happened in other older Indoeuropean conquests, for example in Greece or Italy, or in India too. Meanwhile those who remained behind, established for many centuries in the same place, already considered for one or another reason the women as "their" women and not mere prisoners.

    It's not the nature of the climate but the nature of society what I had in mind: settled down society on one side, recent conquest society on the other.

    ...

    ReplyDelete
  128. "... the Kurgan Model doesn't even confirm anything! Every new find on this field, including genetics, continues to be utterly surprising and contrary to all previous expectations".

    My impression is that you don't understand the Kurgan model, as for example when you claimed that the presence of some West Asian ancestry in Yamna and Corded Ware disproves anything. In fact it is irrelevant to the theory, as it never modeled anything before Samara: the Neolithic origins before that culture were simply unknown. Now we just get to know something more...

    From previous comments:

    "The list of falsified predictions includes the steppe origin of blue eyes (already present in Western Europe long before), lactase persistence (that apparently missed PIE expansion), white skin (already existed in LBK hybrids), YDNA R1a (null in Yamnaya, what a joke!!!), mtDNA H (already found in Mesolithic Iberia)"...

    None of this is part of the Kurgan theory. Just some extrapolations made by some defenders of the demic replacement variant theory.


    "... domesticated horses (actually derived from a western European variety)"...

    You are wrong in this. Horses had two centers of domestication: one in the steppes and the other in or near Iberia. The horses aspect stands, although it's unclear if the IEs were the first ones to domesticate them in the region or rather adopted the idea from some other nearby culture.

    "... absence of IE languages in Bronze Age western Europe (highly hypothetic and contradicted by valid evidence)"

    Absence before the Urnfields expansion is almost certain. Nothing at all demonstrates Western presence of IE languages before the various Celtic or Celtic-plus expansions of the late Bronze and especially Iron Age. This is anyhow very slippery because the linguistic evidence is almost nowhere to be found for those period, so we can only infer from other data such as material archaeology or archaeo-genetics.

    So you're basically just wrong in everything. Sorry, but that's how it is.

    ReplyDelete
  129. 'I think you're whitewashing Germanic mythology.'
    I think you are surprisingly simple minded.

    'how do you explain that the same form exists in Scandinavia, which was not affected by that phase?'
    The Roman border was a hotbed of Germanic change towards an apocalyptic society, this also affected Scandinavia. Moreover, Scandinavia had its own period of devastating anarchy. All of this influenced the contemporary belief system.

    'Each subculture evolved separately and through its own distinctive paths.'
    So how come you feel obliged to generalize on Migration Period Germanic mythology?

    'Romans also had a vast animistic imaginary world. It's very common but it's not the bulk of the religious corpus.'
    This animist layer is indeed very common, it exists in all IE cosmologies. The quantity of Dumezil elements is irrelevant to the coherence and internationality of this deeper layer, that Gimbutas adherents like yourself dismiss as "defeated and subjugated mythology". A quite peculiar stance, really. Especially to somone that claims to know what is "old, intrinsic and very much rooted in the shared Indoeuropean legacy."

    'how do you explain that precisely in the Germanic pantheon we find so very few goddesses, with so little power, and that Valhalla was denied to women'
    You refer to the Norse mythology, that itself is an evolved version of earlier Germanic mythology. The Frisians dedicated their victory on the Romans to Baduhenna, a goddess of war. There were that mythological women such as Hel, Nehalennia, Huld, most likely related to the Frau Holle (Mother Winter) of Grimm's fairy tale. Your view on contemporary Germanic mythology has nothing to do with its original identity and function. World War II may have inspired Gimbutas to think otherwise, but she was wrong. When IE 'invented' their tripartite mythology, at least in NW Europe they were already aware of the triple principle of their gods, that include the mother goddess.Hence, IE culture was not necessarily patriarchal everywhere and down to its earliest stage.

    ReplyDelete
  130. Odin
    Dievs
    Thor

    un
    deux
    trois

    Quite funny indeed!

    ReplyDelete
  131. 'accusing Gimbutas when the newest archaeo-genetic data from Central and Eastern Europe supports her at least in most aspects: there was some sort of major change, even to some extent demographic, in the Chalcolithic precisely with the arrival of Kurgan peoples from the East.'
    You are notoriously vague. 'Some' major change involving people in the Chalcolithic. My horoscope says 'some' important event with a person will happen this week, that's a lot more accurate. And what Kurgan people from the East? If you mean those Yamnaya again, I didn't see them around. True, I saw some genes coming along. These were shared bad016554y the Yamnaya hybrids and must have been from a common source, we still have to figure out exactly where from but I am sure it wasn't exclusive to any typical steppe culture.

    ReplyDelete
  132. 'you don't understand the Kurgan model'
    I can see right throught it!

    'In fact [a Neolithic origin] is irrelevant to the theory'
    Like it would be irrelevant to Out Of Africa that some ancestor of Lucy was the owner of an interstellar shipyard?

    'None of this is part of the Kurgan theory'
    Indeed, the Kurgan theory is a ghost. It can't be measured in any way, how stupid.

    'it's unclear if the IEs were the first ones to domesticate them in the region'
    The Botai people kept herds of steppe horses. You must think that Iberian horses were far superior that IE people prefered exotic horses.

    'Absence [of West European IE languages] before the Urnfields expansion is almost certain.'
    Koch could deduce that Tartessian must have been a Celtic language older than Urnfield. Moreover, that Celtic was a superstratum to an even older substratum of IE languages like Lusitanian, and probably also to a Belgian-like language group that can be traced as far south as the Cévennes mountains. If Luuk de Ligt is right on one of the ‘Eteocretan’ inscriptions from the town of Praisos, then Italic languages are also older than Urnfield. Instead, there is no evidence of absence to support your stance.

    'So you're basically just wrong in everything.'
    Don't be so overtly soft to yourself and go to some greater pains to save your paradigm. How would you prove your falsification criterion true that 'Elite dominance' indeed involved the Kurganic mediation of invisible ghosts?

    ReplyDelete
  133. @Rokus: you are just entrenching yourself in your dogma without providing anything for healthy scientific debate, just authoritarian dismissiveness. Your dogma: your problem.

    ReplyDelete
  134. Also Koch is junk. Anyone studying the matter knows it: he totally reinvents the Ibero-Tartessian script so it fits his whim, his "Celtic" by-product is not too credible anyhow. I was expecting a reference to Koch at any time. In any case Tartessian or sud-Lusitanian inscriptions are all Iron Age and people identified historically as "Celtic" lived nearby, so...

    ReplyDelete
  135. @Maju,
    'authoritarian dismissiveness' and subsequently: 'Also Koch is junk'.
    Look at yourself, are you insane? You urge me to read books of your grandmother, but missed the relevant books of the renowned modern Celticist John T. Koch! Celtic from the Southwest at the Dawn of History, and both 'Celtic from the West' collections, edited by Barry Cunliffe and him. I already had some idea that probably you rarely look in a mirror, but for sure a 'healthy scientific debate' is not your speciality. Indeed, of someone 'entrenching himself in his dogma' you are the living example. As I already said: none of these findings escaped stubborn denials, since Kurganism already has another reality that "coherently explains the expansion of Indoeuropean languages". Unfortunately, that's how pseudo-scientific dogmatic fossils always are.

    ReplyDelete
  136. I've analyzed myself Koch's materials with Untermann's transliterations in hand and Koch makes up almost everything (and, if you use Untermann's method, all or most Tartessian sentences actually appear Vasconic). All linguists, professional and amateurs, with whom I have commented the matter (my "Celtic" is not good at all, so I rely on second opinions for this) agree that Koch hypothesis, widely available online, is simply not credible. How good is your Celtic?, your Vasconic? How familiar are you with the Ibero-Tartessian script? Why do I have to believe you then?

    ReplyDelete
  137. It is misleading to speak of any sort of 'mainstream view' for a subject with so few active researchers. Non-linguistic counter-arguments prevail. Indeed, an Indo-European Tartessian language as a sibling of (q-) Celtiberian remains unacceptable to those that simply deny Bell Beaker being most likely Indo-European. Kurganism still thrives on plain denials of the sort. Koch (2014): 'The reviews also argue forcefully that it is unlikely that the Celtic language family first emerged from Proto-Indo-European in the Atlantic region, or specifically the Iberian Peninsula.'
    Anyway, genetic results already confirmed that Bell Beaker were connected to the IE network. However, I am not sure if Bell Beaker would have been ancestral rather to the non-Celtic IE people of western Europe, the Celts being possibly a more recent 'Thracian-like' addition. Revisionist Kurganists already seem to have come to terms with Koch's views. Mallory was one of the contributers to the second "Celtic of the West" collection of Koch and Cunliffe. what still doesn't solve the origin of the pre-Celtic IE population of the west.

    ReplyDelete
    Replies
    1. Mallory was an invited contributor. He didn't necessarily agree with "CFW" theory. He still fell back to his preferred theorem - steppe/ Kurgan. The major problem of CFW is its isolationist perspective . (a) how did pre-Celtic arrive in Iberia in the first place ? Did it spring out of the ground ? (b) how does it account for other IE languages.

      And Im not sure Koch's interpretation are universally, or even commonly, accepted.

      Delete
  138. moreover, Kochs theorems are wishy washy , and actually pretty weak for a professed linguistic. He insinuates that PIE was some kind of large dialect continuum or sprachbund or convergence area. In fact he does;t define it at all, if he even understands that the above three are all very different types of language contact. And none of them are possible for the scenario he envisages

    ReplyDelete
  139. @Mike Thomas

    Let me throw a trump card that will knock the foothold from underneath Koch's idiosyncratic theory using another radical idea: PIE was a p-less language. Celtic and Armenian, at the two extremes of the PIE expansion from the Steppe, are closer to that original condition than the other dialects. Balto-Slavic also shows some puzzling p-less developments (no reflexes of PIE *pHter 'father' but forms such as *strujus 'father's brother' instead of expected *pHtrujos, *netiji instead of neptiji**, etc.). Elsewhere, p is an overlooked reflex of PIE labiovelar *kw (again, just like in Celtic) and not a reflex of PIE *p. Forms such as Lat coquo (Gk arto-kopos) and quercus that are thought to have lost p- from assimilation from the second -kw- never had p- in the first place.

    If this was proven to be the case, would there still be a reason to have CFW?

    ReplyDelete
  140. 'He insinuates that PIE was some kind of large dialect continuum or sprachbund or convergence area. In fact he does;t define it at all, if he even understands that the above three are all very different types of language contact.'
    One type does not rule out the other, a migration might break up a dialect continuum and change the region into a sprachbund, where convergence is as effective as before. Thus we may have had a continuous process of shared innovations between languages that aren't necessarily from the same stem. One of the contributors describes an innovation shared by Celtic, Slavic and Iranian but not by Indo-Aryan, Italic or Germanic. Suggesting that Celtic was an immigrant language in the west after the 2000 BC breakup of Indo-Iranian languages, and not necessarily from the north as Mallory suggested. Hence, linguists may easily differ on their approach. Kurganists typically vote for an evolutionary tree without much convergence, while normal linguists that are less involved in Kurganist issues, ie. just normal linguistic investigators, rarely tend to downplay the effect of convergence at all. For any living language it tends to be harder to avoid external influences than to stick on their independent evolutionary tree! Especially now the modern world seems to capitalize on their definitely IE heritage and even accelerated the impact of international communication and trade networks.

    'The major problem of CFW is its isolationist perspective'
    From a Steppe stance, certainly. And despite your above suggestion, Koch did not settle on some kind of paleolithic model as a replacement. Instead, Koch invites specialists to think about another model that doesn't necessarily break with other models but rather hosts a more complex narrative. Even if the consequence of this would be that the focus of IE origins should return to the west, you can be sure that Koch would never says so.

    ReplyDelete
  141. German, when I go to the proto-Indo-European Swadesh list at https://en.wiktionary.org/wiki/Appendix:Proto-Indo-European_Swadesh_list
    there are many words that begin with ”p.” What would they be in a p-less proto-Indo-European?
    *pénkʷe = five
    *ph₂tḗr = father
    *pel- = skin
    *pulh₂- = hair
    *pṓds = foot
    *peg-, *psten- = breast
    *pneu- = breathe
    *pleh₂k- = to hit
    *pet- = to fly
    *pers- = dust
    *pent- = road
    *pl̥h₁nós = full

    ReplyDelete
  142. @Kristiina

    Yes, It's assumed that all those p's go to a PIE *p. But then there are labials in such diverse branches as Greek, Italic, Germanic and Celtic in which p's go back to *kw (b to *gw, etc.). Even in Sanskrit there are words such as kloman where k- appears instead of expected p-. Or in Lithuanian IE *pekw- shows up as kepu. In Slavic some words with IE *p turn up p-less reflexes in unmotivated situations. I've argued elsewhere (http://kinshipstudies.org/2014/07/23/indo-european-labiovelars-a-new-look/) that labiovelars may have left a much more complex network of reflexes across IE dialects meaning that even more p's may go back to PIE *kw than presently assumed. From your list *pénkʷe = five shows up with kw- in Italic (quinque) and Celtic (coic). yes, I know, people treat them as assimilational replacement of *p with kw but do we absolutely have to? Without going into all the linguistic nuances of the situation, I'm just curious, if the idea is defensible linguistically, how would it affect the CFW theory?

    ReplyDelete
  143. German,
    I agree with you and find it very strange.
    why they should change from p to kw?.
    BTW where you find Kloman?.
    Another suspect of more original form is Romanian aqua , Skt. apa,
    Romanian aqua is also become apa, the change kw to p is easy, the opposite is quite absurd also Latin coquo Skt pacati,
    I totally support your idea at the moment:)....

    ReplyDelete
  144. @German

    Please be curious, curiosity is the core of science.

    It could also be the opposite: that several daughter languages did not have "p" as an areal feature. Maju would be a better informant but it looks like for example Basque has usually "b", and "p" is not so frequent. Perhaps Neolithic farmer languages lacked "p"?

    In any case, North West Caucasian languages to which PIE is sometimes compared, have often a "p" sound or ps/pš sounds. I see that Hittite and Tocharian reconstructions also have "p".

    ReplyDelete
  145. By the way, Arabic also lacks the "p" sound, and when Persian was written in Arabic script, a sign for "p" sound was introduced.

    ReplyDelete
  146. @Nirjhar

    "BTW where you find Kloman?"

    Skrt kloman– ‘the right lung’, Gk pleumoon ‘lung’, Lat pulmoo ‘lung'

    In Kristiina's list *pneu- 'breathe, sneeze' shows up every now and then with a k- as in Germ. *hneosan 'sneeze' (< *kneosan).

    @Kristiina

    "I see that Hittite and Tocharian reconstructions also have "p".

    yes p is well attested in Indo European languages, so it's not an easy rock to move but still: another item from your list *pHter 'father' is not attested in Anatolian or Balto-Slavic at all and another one *pel- 'skin' lacks a verbal source and does not show up in Anatolian but then Hitt has iskallai 'split, rip' (with other sk- examples all over Indo-European, as in Goth skilja 'butcher') which may provide a natural verbal source for the skin, hide, etc. forms.

    ReplyDelete
  147. Hi German,Kristiina
    About Kloman Mayrhofer says that it is a dissimilation from p because of m...
    I don't think that pneu comes from kwneu, a labial is more appropriate, and Germanic is fneu-: http://www.etymonline.com/index.php?allowed_in_frame=0&search=sneeze&searchmode=none
    The rarity of p in Basque recalls Arabic indeed, although p is present in Proto-Semitic.
    Ah, I forgot that Irish athir is a regular evolution of patr, initial p is lost in Irish (interestingly
    similar to Armenian).

    ReplyDelete
  148. @Nirjhar

    There's no other change in Indic or any other IE language like the one proposed by M. to explain kloman. Hence, it's an ad hoc idea.

    Germanic has both *kneu- and *pneu- reflexes. The whole set is a bit problematic because potentially onomatopoeic.

    "I forgot that Irish athir is a regular evolution of patr, initial p is lost in Irish (interestingly
    similar to Armenian)."

    yes, that was my original observation. But I'm not 100% sure it was *p- in PIE. It's plain p- in Indo-Aryan, Gk, Italic, Tocharian. The root is unattested in Anatolian, Albanian and Balto-Slavic (or attested without p- if Slav *strujus 'father's brother' is adduced). Germanic has f-, which sometimes comes from *kw- (< wolf < *wulhwaz 'wolf' < PIE *ulkwos). h- in Armenian may also come from a laryngeal or s-.

    ReplyDelete
  149. Does it mean then that the words I listed should begin *kw or be borrowed?

    Here are some Eurasian correspondencies with "k" sound:
    *pénkʷe = five, cfr Ket qaγoem, qāŋ, qāk, Itelmen kuvumnuk five
    *ph₂tḗr = father cfr Sanskrit āryaka, Nenets jirikə grandfather
    *pel- = skin, cfr Latin callum, Maltese ġild, Itelmen kilɣi-ɫx skin
    *pulh₂- = hair; cfr Saami guolga, Chuvash khyl, Turkixh kıl, Khalka x́algas hair
    *pṓds = foot; cfr Burusho Yasin hˈuʈi-s foot
    *peg-, *psten- = breast; cfr Albanian gjoks, Lezgi q'ych', Tsakhur koksɨ, Turkish ğöğüs breast
    *pneu- = breathe;
    *pleh₂k- = to hit
    *pet- = to fly; cfr Turkish gitmek to go, Kabardian k'wa-, Korean kada, Yukaghir kie- general motion verb
    *pers- = dust; cfr Finnish karsi dirt
    *pent- = road; cfr Burusho Hunza gan, Saami geaidnu, Ket qÿnäng road
    *pl̥h₁nós = full; cfr Finnish kylläinen full (with food)

    ReplyDelete
  150. … I forgot to add "or onomatopoeic" as German said.
    To me, *kneu- and *pneu- look like imitating the sound of blowing and *pleh₂k and *kleh₂k the sound of hitting.

    ReplyDelete
  151. @Kristiina

    Yes, that would be the expectation. I don't do "Nostratic" comparisons but I did notice that, if reconstructed as *kwe-Hter, the IE term for 'father', would be easily comparable with a solidly reconstructed Uralic set *eka 'older male relative, father, father's brother, etc.' (http://starling.rinet.ru/cgi-bin/etymology.cgisingle=1&basename=%2Fdata%2Furalic%2Furalet&text_number=+133&root=config). If reconstructed as *p-Hter the IE term for father has no external parallels outside of a sea of "baby-babble" words, which are useless for comparative linguistics.

    For the last *plHnos root I did find a plausible k- parallel in IE (http://kinshipstudies.org/2014/07/23/indo-european-labiovelars-a-new-look/).

    BTW, a "labiovelar" is a catchall. It could have been a "palatalized labiovelar" (k'w-) or a "glottalized labiovelar" ('kw).

    ReplyDelete
  152. German.
    I have checked everywhere, there is no *kneu- form, only
    *hneusan and *fneusan. I think that an original kn- is improbable, because it does not express the
    sound of sneezing like hn- or fn-, which maybe does not even come from pneu-.

    About kloman, Mayrhofer does not give parallels, he cites also another linguist who maintained
    that IE *pl/pul- becomes *kl/kul- in Vedic. However, we have plava, plavate... which is the root
    of kloman according to the common view, since in IE languages we have all word from *plau-:
    http://etymonline.com/index.php?allowed_in_frame=0&search=pulmon&searchmode=none
    The idea is that the lung is the floating organ, because of the liquid inside.
    If you check, we cannot find Skt. words with plVm-, so dissimilation seems probable, the fact
    that it is ad hoc does not make it wrong. In Kalasha we have kroma 'lung', implying plauman>klauman>kroma.

    Germanic f in wolf must come from an intermediate wlp-<wlkw--, like in Latin lupus.

    I have no proof that in Armenian h can come from s-, if you see here: https://en.wikipedia.org/wiki/Armenian_language#Indo-European_cognates initial s- in summer is simply lost. From laryngeal it seems is possible.

    ReplyDelete
  153. Like In Armenian there is hoviw 'shepherd', corresponding to Luwian hawi 'sheep'...

    ReplyDelete
  154. Kristiina,
    Nice! but for example we can't compare *pel- = skin, cfr Latin callum, as the root for Latin word is different-
    http://www.etymonline.com/index.php?allowed_in_frame=0&search=callus&searchmode=none

    ReplyDelete
  155. @Nirjhar

    For Germanic *kneu- see: https://en.wiktionary.org/wiki/sneeze. A more in-depth discussion is at https://books.google.com/books?id=plTFCQAAQBAJ&pg=PA15&lpg=PA15&dq=indo+european+*pneu-+breathe&source=bl&ots=7acGLyubPD&sig=-HM6HOWdVBZ9y6L6enkqugqAYUE&hl=en&sa=X&ved=0CDMQ6AEwBGoVChMI25rbkeq8xwIVAVYeCh3xbgTd#v=onepage&q=indo%20european%20*pneu-%20breathe&f=false

    "About kloman, Mayrhofer does not give parallels, he cites also another linguist who maintained
    that IE *pl/pul- becomes *kl/kul- in Vedic. "

    I argued (see link in my my response to Kristiina above) for an opposite development from kw to p using the same material. IE *kwel– ‘full, fertile’: IE *kwel– ‘crowd’ (Skrt kulam ‘herd, lineage’, OIr clan, cland ‘offspring, lineage, clan’, Lith kiltis ‘clan’, Gk telos ‘crowd’, Slav *celedi ‘serfs, servants’) ~ IE *pel– ‘full, offspring, multitude’: Gk pleerees ‘full’, pleethos ‘multitude’, Skrt puurnas ‘full’, Lat pleoo ‘fill up’, pleenus ‘full’, pleebees ‘crowd’, OIr lan ‘full’, Goth full ‘full’, Lith pilnas ‘full’, Slav *pulnu ‘full’, *plemen < *pled-men– ‘tribe’, *plodu ‘fruit, offspring’. There’s a complete semantic and morphological alignment between Slav *celedi ‘serfs, servants’, OIr cland ‘clan’, on the one hand, and Lat pleebees ‘crowd’ and Slav *pled-men ‘tribe’.

    "In Kalasha we have kroma 'lung', implying plauman>klauman>kroma."

    I didn't know the Kalasha form. Thanks! It makes kloman from *pl- even more suspicious. I connected the whole "lung" set with another set that includes Skrt sroni, Gk klunis, Lith slaunis, etc. That's where k- is clearly primary. I doubt the semantic connection between 'lung' and 'float'. The similarity between lungs and buttocks as paired body parts must have occurred to ancient Indo-Europeans as they were cutting up sacrificial animals. Considering s- in sroni we should reconstruct a palatalized labiovelar *k'w-.

    "I have no proof that in Armenian h can come from s-"

    That's a good call. I don't have specific examples at hand, so I need to check into it more. But a passing reference to this change is made on p. 89 in http://www.jolr.ru/files/(128)jlr2013-10(85-138).pdf.

    ReplyDelete
  156. @Nirhar

    "but for example we can't compare *pel- = skin, cfr Latin callum, as the root for Latin word is different-"

    For *(s)pel- I offered *(s)kel- above.

    ReplyDelete
  157. For the very root *pleu- 'swim, float, wash' there's another, similar one *leuh2- 'wash, flow' (Arm loganam, Lat lavo, OIr luaith 'rain', etc.) without p-.

    ReplyDelete
  158. Nirjhar, when I made the list I did not do any further searches, and, in fact, I doubt that pel- is an IE or Nostratic word at all. I cannot check now but IMO there are better cognates for the IE word for skin.

    Also Wikipedia gives for pellis only the following explanation:
    only Cognate with Ancient Greek πέλμα ‎(pélma, “sole of the foot”), Old English fell ‎(“fell, skin, hide; garment of skin”).

    ReplyDelete
  159. Kristiina,
    Pal is IE see the Sumerian list also.
    German,
    I reply on some (I am very busy now),
    On this story of sneezing does not look very interesting ;), about kw- to p-,
    it is the normal evolution, but this does not exclude a dissimilation from pl to kl
    when there is the labial m in Vedic, something that cannot be denied since
    there are no words in Sanskrit with plVm-.
    Armenian s to h should be the intermediate passage before the loss of the consonant,
    as in Ionian Greek.

    ReplyDelete
  160. German,
    Well , the connection of kloman with sroni does not look very convincing, semantically and morphologically.
    The correspondence of Skt. kloman (*klauman) and Greek pleumon (*plauman) with the same meaning
    instead is too perfect to be easily refused. The connection with the root plau/plu is quite necessary if you
    see the other words coming from that root in Greek and Baltic.
    sroni does not come from k'wl- but just k'l-

    ReplyDelete
  161. @Nirjhar

    "The correspondence of Skt. kloman (*klauman) and Greek pleumon (*plauman) with the same meaning
    instead is too perfect to be easily refused."

    Oh, no, I don't doubt the connection between kloman and pleumo:n 'lung', only the connection between IE lung words and *pleu- 'swim'. I just merge IE *klou-ni- and IE *pleu-mon into a single set referring to two similarly looking paired body parts.

    Gk klonis ~ pleumo:n
    Skrt sroni- ~ kloman
    Lith slaunis ~ plauciai
    Lat clunes ~ pulmo:

    "sroni does not come from k'wl- but just k'l-"

    If two sets are treated together, then *k'wl (palatalized labiovelar) can be a protoform for both. k and s are known to alternate in satem languages because they are historically related phonemes, while k and p are not, hence kloman < *ploman is less plausible than kloman < *k'loman.

    ReplyDelete
  162. German,
    On IE *kwel– ‘full, fertile’: IE *kwel– ‘crowd’ i do find that interesting :) but for example
    clan, you know the etymology is made from plant!
    something like Sanskrit
    http://www.etymonline.com/index.php?allowed_in_frame=0&search=clan&searchmode=none

    ReplyDelete
  163. I will reply on others later, its a good conversation:).

    ReplyDelete
  164. "Pal is IE see the Sumerian list also."

    At http://psd.museum.upenn.edu/epsd1/nepsd-frame.html
    I found only the word below so what's your point?

    tug2pala3; tug2pala; tug2pala2 "a garment" Akk. tēdīq bēli; tēdīq bēlti; tēdīq šarri

    ReplyDelete
  165. "but for example clan, you know the etymology is made from plant!"

    That's news to me. Vasmer connects clan with Slav *celedi and Skrt kulam, so I'd need to dig around further. Thanks.

    ReplyDelete
  166. Nirjhar, do you mean that because in Latin there is the word "pallium" which means "cloak" and Sumerian and Akkadian have pala and bēli, respectively, meaning "garment", a word related to Latin pellis is the original word for "skin" in Indo-European?

    Or do you think that Proto-Slavic *poltьno, from Proto-Indo-European *polto- ‎(“cloth”) is relevant here?

    In any case, I think that wowen garments usually have a different etymology from the word "skin" which is a completely different material.

    ReplyDelete
  167. German, I see that you explain the IE side of the root pel- on your blog. 

    However, now that the autosomal analyses have shown that IE languages were spoken by people who had Eurasian hg roots, Caucasian/Teal roots and Neolithic farmer roots, my current view is that all these components should be found in IE languages and not only Steppe hunter gatherer stuff as Davidski would put it. Moreover, Davidski maintains that it was specifically southern women that were incorporated in these societies, so I am sure that they brought part of their lexicon with them.

    Here are again some observations:
    PIE *penkwe–; cfr Estonian “kämble”, (palm of) hand
    PIE *kwekw– ‘bake’; cfr Saami T “kipted”, to cook
    Slav *pleva; cfr Finnish kalvo,, membrane
    PIE *kweH2ur ‘fire’; cfr Finnish “käry”, smell of something that has burnt
    PIE *kwer– ‘oak’: IE *perkwo– ‘oak’; cfr Estonian “kõrb”, genitive “korve”, thick forest

    I could continue, but I have no time. It is in any case interesting if it is in particular these *kw forms that have reflexes in Finnic languages.

    ReplyDelete
  168. @Nirjhar

    "Like In Armenian there is hoviw 'shepherd', corresponding to Luwian hawi 'sheep'.."

    Thinking about another IE root, namely *peH2s- 'protect' (Hitt pahs, Lat pa:sco:, pa:stor, Gk poime:n, Lith piemuo 'shepherd' (sheep and shepherd is a natural semantic pair), Gk pou 'flock of sheep'), it seem to be pretty compatible with Arm hoviw 'shepherd', hawran 'herd' and hence with Gk ois, OIr oi, Lith avis, Hitt hawis 'sheep'. So, we may have one single set here, with p- and p-less forms across ALL IE dialects. Celtic would then reflect a certain PIE phonetic reality and not a post-PIE Atlantic reality.

    @Kristiina

    Thanks. I agree that it's a worthy mental experiment to look for cognates between IE and Uralic languages assuming that IE p-forms could have had a velar onset instead.

    ReplyDelete
  169. @Rokus

    "At least the male elite of Yamnaya society were the exclusively carriers of R1b(xL51), what rules them out as the elite that also dominated roughly related cultures credited for being ancestral to probably all IE cultures except Armenians. To assume that R1a and R1b-L51 should have originated from the lower echelons of Yamnaya society violate both the available evidence and Gimbutas' 'original thing'."

    Although I think the raid / depopulate / expand model is possible the other thought I have is Kurganism in two waves: the first wave being R1b metal workers expanding in all directions in family groups (with the specific R1b clades that exploded in Europe coming from the handful of families lucky enough to have brought LP with them) and the tribal chariot fighters being a later second expansion from regions adjacent to the first (with their first conquest being the original home region of the metal workers on the steppe).

    If correct one possible aspect of this is whenever the metal workers settled in a copper producing region the local culture they settled among may have got a power boost that upset the pre-existing balance of power.

    Evidence for this would be things like Kurgans popping up behind enemy lines (so to speak) and evidence of conflict *after* they arrived rather than before with the cultures centered on copper sources expanding.

    ReplyDelete
  170. Hi Kristiina,
    oops! i meant the pal here-
    http://new-indology.blogspot.in/2015/05/sumerian-and-indo-european-surprising.html
    Sorry not for being specific, anyway, we tried to create an ancient Fashion there;).
    Hi German,
    I give more thoughts on some:
    The couples you have found are very nice, but semantically I do not see a close link between buttocks and lungs,
    besides the fact that they are on the back side (they have very different functions ;)) and morphologically they do not correspond. About the meaning 'floating organ', there is a clever explanation here:
    http://etymonline.com/index.php?allowed_in_frame=0&search=lung&searchmode=none
    and here again: http://etymonline.com/index.php?term=pulmonary&allowed_in_frame=0
    The same is given by Pokorny also.
    We must remember that these organs in ancient times where known more from butchering and cooking than from medical science.For Example in Italian the liver is called 'fegato' from 'ficatum', a word derived from the fig, maybe because of the taste or the use of figs to make the liver of the goose fatter...

    About k'l- to kl- in Sanskrit, surprisingly it exists!!:), it is one of those centum forms neglected by linguists, the root klam- corresponding to śram- 'to become weary'. But k'w in Skt. becomes śv, like aśva from ak'wa.
    About kloman from ploman, you can compare also Irish cland from Latin planta: http://etymonline.com/index.php?term=clan&allowed_in_frame=0

    ReplyDelete
  171. @German
    Yes, I think it is really worth the effort as we know that Finland was part of Corded Ware culture and there is really a very high amount of battle axes found in Finland, including northern part of the country.
    (http://www.eupedia.com/genetics/corded_ware_culture.shtml)

    Moreover, Davidski has shown that Finns and in particular Saamis are genetically close to ancient IE people.

    ReplyDelete
  172. Let me guess, IE substratum in Finnic?.

    ReplyDelete
  173. Ah, I forgot On Slavic root *poltьno,maybe it's related to pal, but maybe its the root of prthu, platys in Greek, 'wide', flat I think in Germanic.

    ReplyDelete
  174. German,
    you have a lively linguistic imagination but that kind of imagination :)
    is one of the big dangers in linguistic research. It's nice to try new hypothesis,
    but with some measure.
    About hoviw, there is also Skt. avi- 'sheep'. Also Toch. B awi 'female sheep'. hawi is not
    Hittite but Luwian. So, it is a specific root for sheep.
    I don't think there can be a connection with the root paH, although this is used for sheep
    keeping. The aspiration of p is found in IE only in Armenian and Celtic as far as I know,
    so I suspect some connection between the two... and R1b can be considered in this context,
    what do you think?:).

    ReplyDelete
  175. @Nirjhar

    Unfortunately, traditional historical linguistics leaves a lot of room for subjectivity. For me, the connection between lungs and flotation is highly fanciful. The connection between lungs and buttocks (and between *pleu- 'swim, float' and *leuH2- 'wash, flow'), on the other hand, is quite logical (lungs and buttocks are body parts, pleu- and leuH2 are verbs related to moving in the water). Reconstruction *pleu- leaves kloman and kroma unexplained. If *kleu-men is compared to *k'leu-n, the problem disappears and the alternation between p- and p-less forms systematically repeats in both the LUNGS-BUTTOCKS set and the SWIM-WASH set. Finally, I think ancient Indo Europeans grouped body parts not in terms of their location in the body but in terms of their similarity to each other (We still say 'cheek' in English for both face cheek and for backside cheek and have no problem doing it.). Hence, lungs and buttocks, knees and elbows, blood and tears were cognitively connected, while the fact that some of them are from are upper body parts, while others are lower body parts, or some of them are internal organs, while others are external organs didn't matter.

    "you have a lively linguistic imagination but that kind of imagination :) is one of the big dangers in linguistic research."

    I completely agree with you, hence I don't really imagine anything. I just follow natural semantic and formal connections instead of postulating "floating lungs" and unique, distant regressive assimilation type of changes, as in kl < pl.

    "I don't think there can be a connection with the root paH, although this is used for sheep
    keeping."

    Even now Arm hoviw is interpreted as *ovi-pa (with -pa- going back to PIE *peh2-), so the connection between the two roots does exist. The question is the nature of this connection. Interestingly, Colarusso, as part of his Pontic hypothesis, compares IE words for sheep with North Caucasian words for graze and uses Arm hoviw as a semantic bridge between the two sets. https://books.google.com/books?id=KDaUezvjKfsC&pg=PA57&lpg=PA57&dq=colarusso+pontic+sheep&source=bl&ots=eC_wliBNsE&sig=qK-1_efv6WgPm2KTt8NwVl5LL4s&hl=en&sa=X&ved=0CCYQ6AEwAWoVChMIoJz-x73ExwIVB_0eCh11NQG5#v=onepage&q=colarusso%20pontic%20sheep&f=false.

    I don't see too many formal problems with postulating zero-grade *pH2ow- for 'sheep' and full-grade *peH2-/ *poH2- for 'protect, graze'. The presence of p- in the sheep words can explain why the front vowel in Gl ois (< *oFis) is rounded. H2 yields /a/ in both Lat pa:sco 'protect', graze' and Lat avis 'sheep'. The bigger issue is whether the p-less onset in the sheep words actually implies that PIE had some kind of labiovelar or voiceless rounded palatal of sorts or it's a cluster of p + H2.

    @Kristiina

    The linguistic connection between Uralic and IE makes a lot of sense of so many levels but so far a few pieces of the puzzle are still missing.

    ReplyDelete
  176. @Nirjhar

    "The aspiration of p is found in IE only in Armenian and Celtic as far as I know,
    so I suspect some connection between the two... and R1b can be considered in this context,
    what do you think?:)."

    As you have noticed, I disagree with the way historical linguists analyze linguistic evidence. I suspect the picture they've painted (including the early split of Hittite and a unique divergence of Celtic) is not accurate. I agree that Armenian and Celtic share p-lessness not through convergence but through shared descent from PIE. But I also think that PIE was a p-less language (and a kw-, xw- kind of language), hence Armenian and Celtic don't share a unique node but show independent retentions of this interesting aspect of PIE phonology. The fact that this trait (p-lessness) has survived at the extremes of IE geographic distribution from a Steppe source (close to the Steppe and far away from the Steppe but not in the Tocharian or Indo-Aryan direction) tells me that the westward IE spread was fast and furious. Did IE bring R1b with them to the Atlantic shore or did they absorb it from a Basque-like substrate? I would very cautiously say that they brought R1b with them and absorbed it from a Basque-like substrate to boot. :)

    ReplyDelete
  177. @German

    "The linguistic connection between Uralic and IE makes a lot of sense of so many levels but so far a few pieces of the puzzle are still missing."

    Maybe I will prove it. :-)

    I agree with you on p-lessness now that I have made these comparisons. It is a good idea. Archaic features tend to be found in the fringe areas.

    I have also been thinking about R1b and Yamnaya, and I have been considering the possibility that their language was more Basque like and not really IE.

    ReplyDelete
  178. @Kristiina

    "I have also been thinking about R1b and Yamnaya, and I have been considering the possibility that their language was more Basque like and not really IE"

    That's indeed a possibility. Will this make Indo-Europeans a post-Yamnaya, R1a-dominant population that admixed with R1b populations in different areas of West Eurasia?

    ReplyDelete
  179. Hi German!,
    I try to be short. From Today I will be around to visit a place for 15 days, so probably will not have time to reply in this beautiful discussion :).
    About subjectivity, when you try to reconstruct the meaning is not avoidable because you have to guess.
    This is a common problem in historical 'sciences'. About formal changes, we can reduce arbitrary subjectivity through the establishment of laws of phonetical change.
    I don't understand which is the similarity between lungs and buttocks ;), the first are also internal organs, so they do not appear like the buttocks. And you also need a root with a meaning to explain those words. Unfortunately the root of klonis and so on does not seem to reveal a more general meaning out of buttocks, but we can search.

    When I spoke of imagination it was about the common root you proposed for Hawi and peH.
    "I don't see too many formal problems with postulating zero-grade *pH2ow- for 'sheep' and full-grade *peH2-/ *poH2- for 'protect, graze'. The presence of p- in the sheep words can explain why the front vowel in Gl ois (< *oFis) is rounded. H2 yields /a/ in both Lat pa:sco 'protect', graze' and Lat avis 'sheep'. The bigger issue is whether the p-less onset in the sheep words actually implies that PIE had some kind of labiovelar or voiceless rounded palatal of sorts or it's a cluster of p + H2. ."

    Now, *pH2ow- seems a normal grade, having o and not zero grade (which is without vowel: https://en.wiktionary.org/wiki/zero-grade). And that the p was lost, even if aspirated, there is no proof of it out of Celtic and Armenian. So, I suggest not to force the evidence to prove a common root that is not possible to trace. When I spoke of connection I meant a common root.
    The vowel o for me comes from the following w, and you should know that I do not believe in H2 giving a, but in the original a. The Caucasian root of grazing is interesting:), and it can be a good etymology for the name of the sheep. I have just read a passage of Ivanova about wool and sheep. In a Maikop site wool has been found. And she says that the earliest evidence for wool is from Shahr i Sokhta I at the end of 4th mill. BC, the same time of the first attestation of 'woolly sheep' in Late Uruk (Sumerian tablets). But already in the first half of the 4th mill. BC we notice more sheep than goats on the Zagros, and more
    animals left alive up to maturity, which is the sign of use of wool.
    BTW i surely think you have something good here :), so i wish you all the best!....

    ReplyDelete
  180. @German

    "Will this make Indo-Europeans a post-Yamnaya, R1a-dominant population that admixed with R1b populations in different areas of West Eurasia"

    Who knows. Disagreement on R1b is rampant.

    In Davidski's first "Smarter than the average bear" graphic, there are indications of this north-south difference, but it is only that graphic. Yamnaya and Afanasievo remain on the same level with Southern Europeans; and Andronovo and Corded Ware are between Yamnaya and extra WHG-high Lithuanian.

    Unfortunately, we cannot infer the language from genes. People with different genes may speak the same language and people with same genes may speak different languages. At the moment I would say that Corded Ware and Andronovo were IE but I am less sure about the others and this is only my partial point of view.

    ReplyDelete
  181. @Nirjhar

    Thanks for all of your thoughts and critical remarks. Some good stuff there! Just short comment to close it off:

    "And that the p was lost, even if aspirated, there is no proof of it out of Celtic and Armenian."

    Arm has hawran 'herd' and hoviw 'shepherd', both with h-, which can be from both p- and H2. Celtic never has h- as a reflex of either *H2 or *p. It just has 0. So your condition is met.

    " And she says that the earliest evidence for wool is from Shahr i Sokhta I at the end of 4th mill. BC, the same time of the first attestation of 'woolly sheep' in Late Uruk (Sumerian tablets). But already in the first half of the 4th mill. BC we notice more sheep than goats on the Zagros, and more
    animals left alive up to maturity, which is the sign of use of wool."

    Some scholars did suggest that IE *ulH2neH2- 'wool' (Hitt hulana) is related to *H2ewis 'sheep'.

    ReplyDelete
  182. Somewhat off-topic, but I'm interested in reading fellow commentators' views (particularly Nirjhar and German's) regarding the Bangani language (if you have any opinions). Looks like it's still an unresolved topic among linguists.

    For unaware readers, it's a northern Indian (Indo-Aryan) language with a layer of seemingly Indo-European Centum words.

    ReplyDelete
  183. "You should know full well that WHG increases affinity to ANE, while SSA and ASI dampen it."

    Not true, ASI increases affinity to ANE depending on proportions. All attempts to cluster ANE show a significant genetic relatedness to ASI, but mostly to west Eurasian.

    Modern euros are like, 90% Western and 2% Indian. Assuming all this Indian is from ANE, and that ANE itself was something like 20% Indian, it means that 10% of the modern Euro genome is ANE.

    Modern north Indians are roughly 35% Western and 60% Indian. Assuming all the Western is from ANE, modern north Indians would then be 41% ANE.

    ReplyDelete
  184. and in the Caucasus who has the peak? Dargins ,Akhvakh?

    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