Update: the paper is now available and open access here.
See also: Yamnaya-related admixture in Bronze Age northern Iberia
...
A paper titled "Ancient genomes link early farmers from Atapuerca in Spain to modern-day Basques", by Günther et al., will appear shortly in PNAS early edition. Here's the press release from Uppsala University.
An international team led by researchers at Uppsala University reports a surprising discovery from the genomes of eight Iberian Stone-Age farmer remains. The analyses revealed that early Iberian farmers are the closest ancestors to modern-day Basques, in contrast previous hypotheses that linked Basques to earlier pre-farming groups.
The team could also demonstrate that farming was brought to Iberia by the same/similar groups that migrated to northern and central Europe and that the incoming farmers admixed with local, Iberian hunter-gather groups, a process that continued for at least 2 millennia.
The study is published today, ahead of print, in the leading scientific journal Proceedings of the National Academy of Sciences of the USA, PNAS.
Most of the previous studies about the transition from small and mobile hunter-gatherer groups to larger and sedentary farming populations have focused on central and northern Europe, however much less in known about how this major event unfolded in Iberia. This time, the research team investigated eight individuals associated with archaeological remains from farming cultures in the El Portalón cave from the well-known Anthropological site Atapuerca in northern Spain.
“The El Portalon cave is a fantastic site with amazing preservation of artefact material,” says Dr. Cristina Valdiosera of Uppsala University and La Trobe University, one of the lead authors.
“Every year we find human and animal bones and artifacts, including stone tools, ceramics, bone artefacts and metal objects, it is like a detailed book of the last 10,000 years, providing a wonderful understanding of this period. The preservation of organic remains is great and this has enabled us to study the genetic material complementing the archaeology,” Dr. Cristina Valdiosera continues.
From these individuals who lived 3,500-5,500 years ago, the authors generated the first genome-wide sequence data from Iberian ancient farmers and observed that these share a similar story to those of central and northern Europe. That is, they originate from a southern wave of expansion, and also admixed with local hunter-gatherer populations and spread agricultural practices through population expansions. The authors noticed that although they share these similarities with other European farmers, this early Iberian population has its own particularities.
“We show that the hunter-gatherer genetic component increases with time during several millennia, which means that later farmers were genetically more similar to hunter-gatherers than their forefathers who brought farming to Europe,” says Dr. Torsten Günther of Uppsala University and one of the lead authors.
“We also see that different farmers mixed with different hunter-gatherer groups across Europe, for example, Iberian farmers mixed with Iberian hunter-gatherers and Scandinavian farmers mixed with Scandinavian hunter-gatherers.” Dr. Cristina Valdiosera adds.
The study also reports that compared to all modern Spanish populations, the El Portalón individuals are genetically most similar to modern-day Basques. Basques have so far – based on their distinct culture, non-indo-European language, but also genetic make-up – been thought of as a population with a long continuity in the area, probably since more than 10,000 years ago.
“Our results show that the Basques trace their ancestry to early farming groups from Iberia, which contradicts previous views of them being a remnant population that trace their ancestry to Mesolithic hunter-gatherer groups,” says Prof. Mattias Jakobsson of Uppsala University, who headed the study.
“The difference between Basques and other Iberian groups is these latter ones show distinct features of admixture from the east and from north Africa.” he continues.
These findings shed light into the demographic processes taking place in Europe and Iberia during the last 5,000 years which highlights the unique opportunities gained from the collaborative work of archaeologists, anthropologists and geneticists in the analysis of ancient DNA.
“One of the great things about working with ancient DNA is that the data obtained is like opening a time capsule. Seeing the similarities between modern Basques and these early farmers directly tells us that Basques remained relatively isolated for the last 5,000 years but not much longer,” says Dr. Torsten Günther.
Source: Ancient genomes link early farmers to Basques
403 comments:
«Oldest ‹Older 201 – 400 of 403 Newer› Newest»Maju
"Bunch of linguistic data". the Vasconic theory has been negatted. hat has been ridiculed Venneman claims to have reconstructed an entire linguistic prehistory based on three toponyms ??
And the EEF came from SE anatolia to northeastern africa . They're genetic homogeneity underlies thousands of years. They'd since diverged linguistically. They arrived as disparate households, clans and tribes. It's absurd to claim they spoke the same language or dialects thereof.
@Mike
"Definitely around the Ponto Caspian, although I'm open to which side (N / S). But the fact is PIE was still unitary circa 2500 BC. So whichever language expanded with the movements of 3000 BC,it wasn't Indo-European"
That is a pretty bold statement. How can you be so sure of the unitary PIE date? If not a mass migration, then what do you think allowed the IE languages to expand so rapidly? A particular technology? Or are you suggesting an additional mass migration of a genetically similar people? I guess I don't follow your line of reasoning. But I am intrigued that someone as logical as yourself has such a take on the evidence.
Venneman's studies are not "pseudo-science". They have been mostly constructively criticized by other linguists, who clearly admit he's onto something particularly with Vasconic (his "Semiditic" was never so persuasive) but that he is committing some lesser errors and also that he may be missing some key data. My own private research is strongly supportive (it's hard to know some Basque and do not realize how many Basque sounding stuff is all around, be it in toponimy or common words) although I also see Vasconic in places like Italy and the Balcans, where he does not, other researchers more focused on this or that have recently also supported quite strongly Vasconic affinities for Paleo-Sardinian and Iberian (and almost the only possible link among these is Cardium Pottery), older researchers like Krutwig already noticed some of that, although like Venneman's work, some was more credible and some probably excessive.
Mostly it's a matter of whether you are familiar with Basque, when you normally see it, or whether you aren't, in which case, most people are dismissive and tend to imagine everything being Indoeuropean. The latter is the typical case of only having a hammer, so everything must be a nail.
Maju
Well I at least partly agree with you in that historical linguistics is need in a major overhaul . It's been waiting since 1923
Mike
I doubt that the age 4500 ybp will be validated any time soon. Heggart's criticism of Haak theory was about the shallow age for PIE (Anatolians included). Removing Yamna out of equation make ages even more shallow. It becomes impossible to explain all the differentiations of branches. Even Chang et al. artificial forcing languages to become young didn't achieve to bring any age younger than 5500 YBP.
lots of gold, silver and copper in Iberia so likely connected by trade to Corsica / Italy-Sicily / Crete
wherever you think it started out from once you get some R1b to Crete I'd have thought it could get to Iberia quite quickly and easily (relatively speaking)
from the frequency it doesn't seem like an invading tribe though more like a small group, family or individual traveling along the pre-existing trade routes among other peoples
.
if some or all of the geographical hops made were male only or mostly male then they could have quite exotic collections of adna depending on the sequence of mothers in possibly different regions
say someone from the steppe arrived near the Pyrenees and married a local woman and their son hopped to Brittany and married a local woman and their son hopped to Ireland and married a local woman and their son to the west coast of Norway etc.
that would need a reason why the hops were male only.
Aram
What you don't know doesn't mean it's not possible.
My figure is a terminus ante quem . . Admittedly.
But it's based on solid linguistic evidence
@Everyone,
An Interesting note.
Of Ancient hg I that has found a subclade beyond I, I2, I2a1, I2a2,.
All of the ancient I clades today are mostly found in NorthWestern Europe except for I2a1a1-M26 which is mostly in SouthWest Europe(especially Sardinia). I2a1b-Dinaric is under I2a1b2 and all the ancient I2a1bs are under I2a1b1 which is most popular in the British Isles. I2a1a2a is also a popular form of P37 in the Isles.
Founder-effects R1b-L11, I1, R1a-Z283 absorbed the older-lineages hg I and IMO it's random the older-lineages hg I have a higher frequency in NW Europe. It doesn't matter if R1b-L11 is of local origin at some point it expanded and absorbed lineages that were once popular.
It makes sense. In East Europe regional founder effects(N1c, I1, R1b-L11, I2a1b2) take up most of their Y DNA that isn't R1a. So old-I lineages are hard to find. In Iberia/SW France R1b dominance and East Mediterranean lineages significantly lower its frequencies.
pre-I2a1a2a and I2a1a2a: Mesolithic Hungary, Neolithic Hungary, Mesolithic Sweden, Neolithic Swedish hunter gatherers.
I2a1a1-M26: Copper Age Italy, Neolithic France.
pre-I2a1b: Mesolithic Luxmborug and Mesolithic Sweden.
I2a1b1-L161.1: Neolithic Spain, Neolithic Sweden, Neolithic Germany, Neolithic Swedish hunter gatherers.
I2a2a1-CTS616: Megalithic Spain.
I2a2a1a2a2: Bronze age Hungary
I2a2a1b1b2: Yamnaya
I2a2a2-Y6098: Copper age Spain
I2c2: Mesolithic Sweden.
I1 and pre-I1: Mesolithic Sweden, Neolithic Hungary.
I may well not understand this right but...
would diversity be affected by effective population size?
if a population moved from the cold low population north to the warmer higher population south might not the diversity increase faster in the south?
Mike
The last 50 years we moved where we reached now based on solid linguistic/archaeological evidences. I respect Your knowledge in many fields, but I will continue to doubt that there is a some solid linguistic evidence out there that is capable to explain all the complexity of IE family. And all that taking genetic data into account.
@Aram "On what haplogroup You will put Your money for Anatolian IE speakers ?"
May be R1a. Turks have quite a lot of R1a: Istanbul 9%, Dogu Anadolu 11%, Guneydogu Anadolu 14%, Akdeniz 9% (these are the highest frequencies; I could not find the original paper and these are from http://www.theapricity.com/forum/archive/index.php/t-135310.html). The average is between 5%-12%.
Mike, it is a good point that today, Yamnaya-related ancestry proportions in Europe and west Asia are concentrated in non-IE's: Udmurt, Mordovian, Bashkir, Volga Tatar, Lithuanian, Pamir Tajik. http://eurogenes.blogspot.be/2015/03/yamnaya-related-ancestry-proportions-in.html
In any case, it is apparent from the R1b trees that European R1b (L51) is not derived from Yamnaya R1b which gave rise to Z2103.
Maju, why on earth Khanty and Nenets should be the original proto-Uralic speakers. Both groups have mixed extensively with Kets, and the title of the recent paper was "Genomic study of the Ket: a Paleo-Eskimo-related ethnic group with significant ancient North Eurasian ancestry".
Kristiina
You're on the ball. Some just choose to close their eyes .
Aram
Thanks for the nice words. Sorry but I can't elaborate.
But be assured my shoulder will be there for you to lean on when the "genetic trail" proves far more complex than espoused, and it comes bleedingly obvious that there wasn't a "big-bang " from the Volga-Don. :)
When one further considers that Indo-european archaeolgy is a construct, the entire house of cards will collapse. And someone Who knows what they're talking about will need to step up.
@Kristiina
Now that Mike Thomas has clearly expressed his thoughts on the issue (although he continues to withhold secrets too seemingly silly or too dangerous to share) would you like to do the same? Who do you think brought IE languages to Europe? Mike says it was after these Yamnaya-like steppe people admixed all over the place. What do you think? Earlier or later?
@Karl K
Corded Ware (2900–2450/2350 cal. BC) fits quite nicely with IE's, and also Sintashta (c. 2800–1600 BC). In Europe, Urnfield culture (c. 1300 BC – 750 BC) looks like it could be the origin of Celtic languages. For the lack of any relevant yDNA, it is difficult to say anything about Hittites (1600 BC) and Tocharians (6th to 8th centuries AD).
@Kristina
Lithuanian, and Pamir Tajik are IEs.
@Kristiina
Ok. Makes sense to me. We are in somewhat of an agreement. It is only that you do not think that Yamnaya was the linguistic ancestor of Corded Ware. Correct?
I think this differs from what Mike believes.
@Mike
"Why do we maintain that IE derived from CWC/ Yamnaya when even the easternmost European IEs do not cluster with them. Yet non-IE do. So we need "Indo EUropean" yamnaya like populations maintaining their language after centuries of admixture with different groups - not to mention - significant social and political changes - yet groups most genetically 'faithful' to these Yamnaya groups somehow language -shifted."
Great question! Seriously. Now we are at the real issue.
I simply see this eastern migration from the steppe with a large impact on eastern european genetics. The R1a territory if that's your thing. This has to be from Corded Ware because they are the ones actually tested and sequenced. There is no controversy about this. Their ancestors were similar to the Yamnaya. Obviously. Why were they able to move in in such large numbers? I don't know.
From the west, I see something similar but different. Some kind of Yamnaya related people but probably with a different culture. If they made it west first then either they were extremely mobile using horses or seafaring or both. Some of them were lactase persistant and this trait was utilized in the milking communities of the Neolithic in Iberia and up the Atlantic Coast and into Britian. Because of their mobility and technology, they also made linguistic impacts throughout Europe. This could muddle the lines of language origin.
If two different waves made a similar impact on language and genetics, then I would say they are related. As for the other regions, sure. Greece is strange and complicated. Anatolia is strange and complicated. India is strange and complicated. Most of the complication is from a lack of information. Bt it doesn't change the basic facts.
Where and when was the PIE language/culture? Not what I am discussing here. Who brought IE to Europe? This is the topic at hand.
Karl
There is no doubt of the demic movements of R1a and R1b.
R1a, IMO, is more straightforward. This was essentially a relative drag nach westen- filling in spaces in central -eastern Europe.
R1b is too complex for me, a.t.m. Lets wait and see what we unfurl in near future.
Let's clarify these issues, then we can talk language :)
But just consider that language expansion can occur without 'folk-migrations'.
In fact, the Copper Age CWC & BB cultures left no real political legacy, and were'nt real Chiefdoms.
L51(xL11) or L51* is found at its highest in France and northern/central Italy...
http://secherbernard.blog.free.fr/public/.L51_Map_with_Neolithic_Path_003_m.jpg
The highest single frequency ever found is in Tyrol, Austria on the Italian border. However, its distribution is really not ancestral to anything, as it is made up entirely of a now well known brother clade of L11 known as Z2115. Based on STRs, it looks to be either as old as, or maybe even slightly younger than L11. If I had to guess, it may have traveled with U152.
"In fact, the Copper Age CWC & BB cultures left no real political legacy, and were'nt real Chiefdoms."
I agree whole heartedly. But they did leave a very strong genetic legacy. How can this be denied?
I think language is perpetuated through close family. But it expands with economy and culture. Sometimes there is a case where these coincide. Native American languages could go in a million directions, because there were no people in the new places.
In Bronze Age Europe, there were language expansions because of technology. And, there were genetic expansions because of technology. These were probably the same technologies.
@Mike
Ok. Interesting.
I actually disagree with institutions and power driving language change. I think more along the lines of high mobility, economic freedom, and new esssential technology. Maybe only ' @Grey ' seems to understand this.
But in this case we likely have a genetic component to the new essential technology. This makes it not one issue or the other. They are intertwined.
Karl
Its not power and institutions only which drive language change, in fact Ive always stated that language change occurs due to many things. Power and institutions secure continuity and making them permanent, rather than just easily changed or absorbed.
Kristiina
The majority of Turkish R1a is Z93, some of them from Iranic period, some a recent Turkish expansion like Kyrgyz case. There are also many R1a-Z93 of Kurdish origin.
I don't know can Anatolian languages derive from Sintashta.
You can take a look at this maps for various R1a subclades.
http://www.eupedia.com/europe/Haplogroup_R1a_Y-DNA.shtml
Btw Looking at that map You can see that not only Anatolian is problematic to derive from R1a but also the Latin.
@Mike
Cool. Totally in agreement then. I see where you are coming from. I have no idea how your thoughts disagree with my own, yet we come out with different histories. I guess this is why there are such heated debates in the field.
Mike
Thanks for the shoulder. But I don't think I will need it. I am not a linguist who invested all his life into Indo-European studies. I am just science consumer. :)
BTW the Udmurts who are considered to be the most faithful ancestors of Yamna, have more than 60% of N haplogroup and and some 20% of Siberian admixture that was completely absent from Yamna. Just FYI.
Udmurts aren't the most similar people to Yamnaya overall, they just have the most Yamnaya component in certain ADMIXTURE runs. The likes of Lithuanians and Mordovians are closest in the overall sense, depending on whether you use f3, fst or whatever.
Barring ancient DNA evidence to the contrary I see no reason to consider Udmurts Proto-IE or even Proto-Uralic although their geographical location matches, since such a thing would require 4000+ years of genetic continuity. The area now is a patchwork of different peoples, Perm and Baimakskiy Bashkirs have over 90% R1a and R1b, and even more Siberian+East Asian than Udmurts, or turkics like Volga Tatars and whatnot.
@Aram “not only Anatolian is problematic to derive from R1a but also the Latin”
Indeed. If we want to derive entire language families from one haplogroup, we are in trouble; but it is also impossible to reserve old and widely distributed haplogroups such as R1b and R1a to only one language family. That is madness.
Swedes who according to Davidski’s calculations are 47% Yamnaya and 47% pre-Yamnaya and 42% hg I and 16% R1a and Serbs who are are 34% Yamnaya and 45% pre-Yamnaya and are 42% hg I and 16% R1a are said to be direct discendants of Yamnaya. Udmurts who are 65% Yamnaya and 16% ENA and 14% pre-Yamnaya and are 87% hg N and 10% R1a, and Erzya who are 53% Yamnaya and 6% ENA and 32% pre-Yamnaya and are 40% R1a and 20% N are said to be a different stock because they speak Uralic languages that in reality are full of all kinds of lexical and structural similarities with IE-languages. Compared to Udmurt non-Yamnaya portion (16% ENA, 14% pre-Yamnaya), Swedes and Serbs have more “foreign blood” and also the R1a percentage of Swedes and Serbs is lower than that of Erzya.
I also checked R1a subclade distribution in Turkey according to Underhill: in Cappadocia and eastern regions there is 0.5-2% of some old M420(x SRY10831.2). In eastern regions there is also 0.5% M198(xM417) and 1.9% Z282. In all regions there are small amounts of M458, 0.6%, and M558, ranging between 0.5-3.7%. The rest is Z93 and derivatives, 1.8% (West), 3.4 (Central), 5.3 (East). It would be interesting to get Hittite yDNA; I would expect it to consists of several haplogroups, including R1b.
Just for the fun of it using Jtest on a 5000-/+ year old R1b Yamnaya sample. And make a tweeked 5000 year old version from same area.
Admix Results (sorted):
# Population Percent
1 EAST_EURO 34.16
2 NORTH-CENTRAL_EURO 18.2
3 WEST_ASIAN 17.74
4 SOUTH_BALTIC 17.69
5 ATLANTIC 9.52
6 SOUTH_ASIAN 2.68
Just for the fun of it if I would make a synthetic R1b sample from the same region- and would tweek it like this, higher Ashkenazi score and lower West Asian.
# Population Percent
1 SOUTH_BALTIC 29.08
2 EAST_EURO 25.02
3 NORTH-CENTRAL_EURO 19.31
4 ATLANTIC 12.2
5 ASHKENAZI 3.24
6 WEST_ASIAN 2.35
7 EAST_MED 2.21
8 SOUTH_ASIAN 2.02
9 EAST_ASIAN 1.55
10 WEST_MED 1.53
11 SIBERIAN 1.44
12 WEST_AFRICAN 0.06
I could do the same with Gedrosia k11
Admix Results (sorted):
# Population Percent
1 WHG 50.78
2 Caucasus 25.02
3 Gedrosian 16.02
4 Kalash 6.53
5 S_Indian 1.2
6 EEF 0.44
Tweeked synthetic modern sample might look something like this.
# Population Percent
1 WHG 49
2 Caucasus 20.64
3 EEF 13.83
4 Gedrosian 5.39
5 Kalash 4.17
6 E_Asian 2.26
7 Indo_Chinese 1.53
8 SW_Asian 0.97
9 Siberian 0.96
10 S_Indian 0.68
11 SE_Asian 0.58
1- WHG (W European Hunter Gatherer) - Loushbour & NE Europeans
2- S Indian - Various S Indian tribal populations, such as Hakkipikki and Nihali
3- Gedrosian - The Baloch, Brahui, and Makrani of Pakistan
4- SW_Asian - Saudis, Yemenis, and Bedouin
5- Siberian - Nganasans
6- EEF ( Early European Farmers) - LBK, Sardinians, and Stuttgart
7- E Asian - Ulchis
8- Caucasus - Georgian, Abkhasians, Adygei, and Balkar
9- Kalash - Kalash of Pakistan
10- Indo-Chinese - Kusunda peoples
11- SE Asian - Ami & Dai.
@Kristina
Slightly tangential to the topic, but related to something that Maju said regarding seeing Basque-related influences in many places: I remember that there were some studies (don't recall now, that's from my university times 20 years ago) about Basque phonetic influence in Spanish. But I have the feeling that there might be more to it than just Basque -> Spanish. There seems to be a common phonetic influence in Southern European languages independently of the family they belong to. Do you know if there is any study regarding this? Comparing phonetic similarities and differences between southern vs. northern European languages?
@Kristinna,
Those Y DNA frequencies are because of founder effects.
"I actually disagree with institutions and power driving language change. I think more along the lines of high mobility, economic freedom, and new esssential technology."
https://en.wikipedia.org/wiki/Cyclops
"They provide Zeus' thunderbolt, Hades' helmet of invisibility, and Poseidon's trident, and the gods use these weapons to defeat the Titans."
"Collectively they eventually became synonyms for brute strength and power, and their name was invoked in connection with massive masonry. They were often pictured at their forge."
"These Cyclopes also created Poseidon's trident, Artemis' bow and arrows of moonlight, Apollo's bow and arrows of sun rays, and Hades' helmet of darkness that was given to Perseus on his quest to kill Medusa."
Just a thought.
@Mike: "It's been waiting since 1923".
I read that at first as "I've been waiting since 1923" so I was picturing you like some sort of mythical druid, Merlin style, lurking from the depths of the past and awaiting for something to happen... XD
...
"Hittite is a red-herring. Marked substratum and areal effects make it appear its much older than it is".
The Anatolian branch overall is probably quite old anyhow, at least if we accept that it crossed the Caucasus range with the Kura-Araxes culture, one of the oldest Kurgan offshoots.
@Grey: FYI Iberian trade connections with the Eastern Mediterranean are now solidly documented since some point in the Chalcolithic. The "colonial hypothesis" about the birth of the first Iberian civilizations c. 2600 BCE (?, chronologies get revised all the time) is still floating around (not without problems) but, as you say, it does not look like they were a huge bunch of people immigrating either.
...
"would diversity be affected by effective population size?"
Absolutely. Everything else equal, drift will always be more intense in smaller populations, reducing diversity.
"if a population moved from the cold low population north to the warmer higher population south might not the diversity increase faster in the south?"
But only from the baseline that the migrant population already carried and subject to the relative slowness of coalescence of novel mutations. In general terms a N→S migration with 100% replacement (hypothetical, of course) should leave a signature of greatly reduced diversity because that's what they probably carried from the North, where their Ne was smaller.
But assuming 100% replacements is almost certainly out of the realistic scope, more so considering that the numbers the North could provide before the Medieval agricultural revolution (heavy plough) were usually very small.
Aram
How's your story of R1b and its 'Indo -european roots'
coming along ?
@Richard Rocca: According to that map, L51* (aka M412*) is concentrated around the Massif Central, with a principal branch towards north Italy and Sicily that reminds a bit to the distribution of U152 (a distant subclade via S116) or less strictly to the general spread of S116 itself.
If that's correct, then it seems pretty much supportive of my notion of Western R1b spreading from roughly "France", and not just for S116 as was my argument so far. That's pretty nice. Do we have more data on this matter?
@Alberto: "Basque phonetic influence in Spanish".
Of course, Castilian (Spanish) is strongly Basque-influenced, but that's a Medieval thing. Just consider words like chico/chica (← [eus]txiki = small, often used for kids and youngsters, and sometimes sticks as nickname for all life) or cerdo (← [eus]zerri = pig). After all the first Castilian is documented at Valpuesta right at the modern border between Castile and the Basque Country. Castilian falls in this sense in the loose category of Vasco-Romances, along with Navarrese-Aragonese and Gascon, although it also has a strong influence from Leonese/Asturian, close to Galician/Portuguese, naturally. When you travel through Spain or look at maps you see a zillion Basque-like toponyms, most of which must be real and also old. However are they Basque, are they Iberian, are they "Tartessian"? We can't know.
But it does not just happen in Iberia, but all around Europe. Adur river in England and Adur river in the Basque country? Arezzo in Tuscany and Areso in Navarre? A zillion ur-/dur-/tur- river names (ur means water, adur salive but also the magic flow, similar to medieval "humors"), also iz- (yss-) ones, from an ancient iz- root also meaning water, fossilized in words like izotz = ice (oops another vasconic word in Germanic: ice!), or itsaso = sea (another one: "sea" is not IE, and looks like *iz-aso could be a relative again, and maybe also of non-IE Greek thalassos), or izurde (dolphin), etc. Or look at Iber (Ebro), Tiber and Ibar, all around the Mediterranean, "ibar" meaning river bank and clearly Basque (relatives: ibai = river, ibon = creek and possibly ibili = to walk, maybe originally also perceived as "to flow"), an a-/e- Basque/Iberian shift is apparent in other words like the copulative particle ke (Iberian), *ka (ancient Basque, now only fossil). Etcetera, etcetera, etcetera...
But you need to know some Basque to perceive all this. If you don't (and most linguists do not) then it's just gibberish.
Maju
"Absolutely. Everything else equal, drift will always be more intense in smaller populations, reducing diversity."
ty
Maju
Be optimistic mate. There's a massive diachronic study in Iberia, they're testing the Balkans, Greece and the Black Sea littoral. I trust we'll still have a few marbles left in 3- 5 years time.
Saying that Corded Ware don't cluster with modern Indo-European speakers misses the point.
Hundreds of millions of people around the world of Northern and Eastern European origin who speak Indo-European languages can be modeled as mostly of Corded Ware origin.
Also, claiming that Corded Ware, and their immediate descendants like Sintashta, weren't Indo-European speakers simply isn't parsimonious, and that's an understatement. That's because there's no other way to get Indo-European languages and R1a-Z645 derived lineages to both Eastern Europe and South Asia.
Dave:
Sure, they can be modelled as descended, in part/ mostly, from CWC. Bue the simple fact is yamnaya and even CWC today do not cluster with Europeans IE speakers. So if we're going to follow genetics, then CWC and Yamnaya spoke something akin to Caucasian or Uralic, or some other idiom. Unless, again, we invoke the idea that those CWC-derived people who did admix in EUrope retained their language, whilst those that in the northeast shifted their language to Uralic. Is that parasimony ?
"That's because there's no other way to get Indo-European languages and R1a-Z645 derived lineages to both Eastern Europe and South Asia."
Actually there is if one doesn;t subscribe to simplistic models of haplogroup = language. The Copper age was the peak of language diversity. It goes against everything we know about sociolinguistics to claim that all R1a-M417 people, or all CWC people spoke the same language. Sure, the R1a trail from EE to India demands explanation. But IMO the evidence at hand already goes against the classic Kurgan model.
You're not making any sense.
Corded Ware only differ from modern Indo-European speakers of Europe in terms of ancestry proportions. However, they carry the same ancestry components.
On the other hand, Uralic speakers carry a Siberian component that is missing in Corded Ware. In fact, it's this Siberian component that drags some of the Uralic speakers east to make them look like they generally cluster closer to Corded Ware in some PCA dimensions.
Considering this, your claim that Corded Ware may have spoken Uralic languages is based on what exactly?
And Caucasian speakers don't cluster closer to Corded Ware than Indo-European speakers do, so they're also irrelevant.
COrded Ware aren't the direct ancestors of Germanic and Balto-Slavic groups. The advent of Germanic and balto-Slavic groups were later phenomena, which occurred in an R1a -rich part of Europe. However, those phenomena had little a priori to do with the earliest arrival of R1a.
So we can accept that Sintashta c. 1800 BC spoke IE, and invaded Asia. But this does not back-prove that CWC groups from 2700 BC in Denmark spoke IE. In any case, central Asia also shows links with other parts of Eurasia (namely west Asia), and it even shows evidence of diffusion out of Asia.
"your claim that Corded Ware may have spoken Uralic languages is based on what exactly?"
based on knowledge of sociolinguistics and lack of bias.:)
Several languages were likely spoken by CWC. I'm using "Uralic" and "Caucasian' in broad terms. Some of the languages spoken are probably now extinct.
You need to understand a couple of points for future reference:
- Corded Ware and Yamnaya show the highest genetic affinity to present-day Indo-European speakers from Northern and Eastern Europe.
- in dimension 1 as well as multiple dimensions of genetic spatial variation, Corded Ware and even Yamnaya cluster closer to present-day Indo-European speakers from Northern and Eastern Europe than to non-Indo-European speakers.
http://eurogenes.blogspot.com.au/2015/09/a-multidimensional-approach.html
- Uralic speakers are genetically highly variable and admixed groups, and most don't show any special relationship to Corded Ware or Yamnaya, unless they look like they clearly have Indo-European ancestry.
So there's absolutely no reason for you to claim that Corded Were were Uralic and not Indo-European, or even both Uralic and Indo-European rather than just Indo-European.
If you think Corded Ware spoke some extinct non-Indo-European language, then you'll have to try and prove that with something else than genetics, because genetics also doesn't back your case in this.
Dave
Genetics doesn'talways equate with language. And a population doesn't need to become extinct for its language to Change. Copper age systems were volatile, and liable to shifts based on factors other than mass migration.
So PIE was Probably one language spoken in CWC, but not the only or even dominant. The ascendency of IE languages in Europe was a later ohenomenon, and cannot be simlplistically linked to the "coming of the corded ware people".
The Yamnaya are totally irrelevant. I'm not denying a connection between R1 peoples and Indo European languages, but wherever they came from had nothing to do with the Yamnaya. Yamnaya were just an offshoot that fizzled out with little impact beyond their bronze age range, we can see this by the distribution of Z2103.
@Mike
"The ascendency of IE languages in Europe was a later ohenomenon, and cannot be simlplistically linked to the "coming of the corded ware people"
"But what does this crazy person suggests caused IE to spread in 2500 BC ? There was no "invasions". This is where real knowledge and insight comes to play ..."
So tell us Mike! We are all dying to know what drove the IE spread.
Seriously. Because I am having a hard time following your subtle logic on this point.
@Mike,
"COrded Ware aren't the direct ancestors of Germanic and Balto-Slavic groups."
I agree with the statement "may have not been the direct ancestors" but I dis agree with saying "there's no way".
@Romulus
"I'm not denying a connection between R1 peoples and Indo European languages, but wherever they came from had nothing to do with the Yamnaya."
Nothing?
Could you clarify. Perhaps the Yamnaya were a dead end, but they were at least closely related to the people who were the big game changers. This is not nothing. Thia gives us a lot of useful information.
@Mike,
"The ascendency of IE languages in Europe was a later ohenomenon"
The specific languages that existed in the Iron age and modern times spread recently, but their ancestors had been in the regions for 1,000s of years. There's no prove that for example Slavic languages expanded over non-IEs. They likely expanded over already existing IE-speakers.
If it wasn't for early writing we would have no idea the Gauls, CeltIberians, Ligurians, Illyrians, Italics Thracians, etc. existed. Some would say IE languages expanded after 0AD with Latin, German, and Slavic. Before Thracian, Ligurian, etc. expanded there could have been IE languages in those regions that went extinct and were never recorded in writing.
@Grey
"Collectively they eventually became synonyms for brute strength and power, and their name was invoked in connection with massive masonry. They were often pictured at their forge."
They name these periods after different metals for a reason. It is a long lasting indicator of technological innovation. And technology spreads through language. In the past, new non-farming technology spread widely and quickly by mostly men. And their sons who learned the language associated with that technology were the ones to spread it.
Why is English so useful today? Not because powerful people force it on other cultures (although that has happened). It is because this has become the language of modern science, computers, and entertainment. It is useful to know it. It helps you bridge cultures when people know a third common language.
Mike
I have a friend who has a good knowledge of this linguistic/archaeological stuff. So what I will say is based on his opinion plus my own genetic knowledge.
There is absolutely no reason to doubt that Yamna is Indo-European. It has the needed admixture components that was predicted based on the peculiarity of PIE. You can reread Bomhard what makes PIE different from historic West Asian and other North Eurasian languages. This peculiarity isn't easy to explain. For linguistic case it is in the right place, having contacts both with Southern languages and Uralic in the North.
One of his branches Z2103 nicely help to understand the origins of Latin ( in South Italy it is much higher than in North ), Anatolian languages ( maybe they have some extra lineages) and Armenian. Also probably Greek, but I suspect Greek will have some mix situation. Also probably some Paleobalcanic languages like Dacian and Thracian.
About the origins of L23 itself. It is possible that it had a southern origin (South Caucasus - North Iran), but it is possible that it had a Northern one. I don't know yet. But even if it had southern origin this doesn't mean that IE was spoken in the South, and a 'teal' alone is a IE marker. No. The formation of IE is a synthesis from 'teal' - South and EHG - North. And the further distribution of IE shows that logic. Any nation who is very high in 'teal' alone is not IE , any nation who has very low 'teal' is not IE, and some nations who are very high in EHG with some extra Siberian/East Asian components show a linguistic shift. If needed it is even possible to make a list of Eurasian nations where all that logic will be clearly visible. They are also exceptional cases like Armenian who is very low in EHG, but we already knew from ancient DNA that the situation was different in the past. There is also the case of Indian languages with low EHG. I expect some surprises from there, but this question is discussed so much that we just need to wait for aDNA.
For Corded Ware I see no real problem. It is possible that the Western part of Corded Ware is at the origin of Germanic languages who came in contact with mysterious L51.
About the West European L51. Some months ago I expressed doubts in Facebook that the R1b-L51 is not IE. There was something strange in it's internal distribution also the Basque/Iberian problem. And I also added the possible Pict issue in Scotland. Many people tried to convince me that I am wrong. So I decided not to pressure to much that question.
What are the origins of L51? Maybe from Carpathian region. My friend proposes to search them in Ezero culture. Usatovo culture. The culture Brno-Lishni. But also as South as Troy. But it is possible we will not find them there and it moved to West Europe from West Asia by a Southern route. Then out of Iberia, out of Italy, France and even out of Britain is possible.
Then the question how L51 relates to the IE. I think it is very short-sighted to claim that if L51 is not IE then it has nothing to do with IE in West Europe. Something happened in West Europe some 5000 years ago. And this something did have dramatic consequences. I would say that the apparition of L51 predisposed West Europe to cultural and linguistic changes. Otherwise many Neolithic people like Basques could survive in vast areas.
One of the theory is the big epidemy that decimated the native population. R1b could have some advantage related to his genetics and origins. I can expand that question further there are historic analogies. A major epidemic can explain the mega founder effect that resulted. From a study on Y DNA bottlenecks we knew that the Europe was the most affected. In contrary Greece and Turkey had very little demographic crisis in this period. This is a just theory. there could be other reasons also.
This is the graph of Effective population size in Europe.
http://4.bp.blogspot.com/-m4-XGZ-zFoY/VVzXWxtK9aI/AAAAAAAAKF0/XELjAeB6_54/s1600/ncomms8152-f2.jpg
Now look Turkey and Palestine are not affected. The impact on Greece is small. Saami declines and don't rebound.
The most dramatic events happen at West Europe especially in South-Western part. There is a big fall and further rapid expansion. England is also heavily affected.
Of course this event is not directly related to the IE question. But as I said most probably this predisposed West Europe to a major linguistic shift. That happened little bit later.
p.s. Btw if someone thinks it is personal for me because I am R1b, I will say that I don't know yet my haplogroup but my predicted haplogroup is J1 based on my relatives. :)
Most of Europe was Indo-Europeanized during the Late Neolithic-Bronze Age period (LN-BA). This is also the time when the modern European gene pool formed.
Ancient and modern genetic data, when interpreted correctly, shows this clearly.
https://drive.google.com/file/d/0B9o3EYTdM8lQX0djRko2WTNkRGM/view?usp=sharing
Moreover, it's very easy to implicate Bronze Age steppe populations in this process, using DNA, archaeology and linguistics.
What happened after the LN-BA in Europe is important to the current linguistic landscape there, but not relevant to the initial spread of Indo-European culture across the continent.
@ Krefter
I don;t think Ive said there is "no way" that CWC is IE (and if I ever did I retract it). Im suggesting that it isn;t necessarily, or perhaps better, not all of it spoke IE - and moreover, not yet the specific pre-runners of modern central European IE languages (ie Celtic, Slavic, German ic).
BTW,
If you've had your mtDNA tested please email me your data(sammyisaac107@gmail.com) or post it here.
I have a User mtDNA spreadsheet here.
https://docs.google.com/spreadsheets/d/18sDtDhgzHScP2_vO6NlfR2rBqefF78JE6bYHZO39WNA/edit#gid=0
After I gather data from Ian Logan mtDNA trees will be better defined and you'll probably have close matches there if not in Anicnet/modern data I have saved.
Aram
"Then the question how L51 relates to the IE. I think it is very short-sighted to claim that if L51 is not IE then it has nothing to do with IE in West Europe"
Of course L51 has eventually very much to do with IE, but maybe not from the outset, and certainly not from the M297 stage.
@mike
"When one further considers that Indo-european archaeolgy is a construct, the entire house of cards will collapse."
I want to clarify this issue and bring about some common ground.
According to linguistics, PIE was a spoken language and so it originated in a place and time.
This ancestral language evolved into multiple languages across time and space resulting in a tree like relation IE languages that we see today and in the historical record.
According to archeology and linguistics the best candidate we have for the (late) PIE speakers are the yamnaya, which is a version of the kurgan hypothesis.
Here is where the dna evidence comes into play.
R1a-M417 and R1b-L23 are great makers for IE because together they spread over the modern (pre columbian) IE world and pretty much only the IE world, with the exception of areas overrun by turks and uralics.
Autosomal dna is a bit more tricky in this context because geography, rather than ethnicity, tends to play a much bigger role. For example, yamnaya is very close to Uralic in autosomal dna but not ethnically/linguistically. HOWEVER, as you move away from central asia the yamnaya component becomes a great predictor for IE influence, presumably due to IE migrations out of the steppe region.
Now, the real kicker. The yamnaya (or a similar steppe group) were surmised as being late PIE on the basis of archeology and linguistics. Also on the basis of linguistics and archeology, IE is supposed to have spread west and east during the metal ages. INDEPENDENT of this hypothesis we now have clear evidence for a yamnaya/CWC like migration moving in the exact locations and times predicted by the kurgan hypothesis.
What we have are two converging lines of evidence telling stories that have a lot of overlap, rather than a relation between the two due to cyclic thinking.
Ultimately IE is a linguistic question. However, given a minimal amount of linguistic parameters I think we would be safe in saying that the yamnaya like group gave europeans the IE language.
Yamnaya culture is NOT the first IE (or Kurgan) culture. It is the local descendant of older cultures in the same area: Khvalynsk (which also occupied the Maykop area of North Caucasus) and Samara, which was apparently quite smaller.
By the time Yamnaya appears, IE was already in full expansion: in Europe west of the Don (Sredny-Stog II and other groups), as in Siberia (Afanasevo) and West Asia (Kura-Araxes). This explains why Tocharian and Anatolian always seem the oldest diverging branches (European groups would still interact for some time but these two became basically isolated very early on).
Where did you get this?
Yamnaya -> Catacomb -> Andronovo -> central Asia
Here's a map drawn up before all the ancient DNA came in.
http://tinypic.com/view.php?pic=157ejye&s=8
The origin of Catacomb culture has always been confusing. It's clearly "kurgan" but the exact genesis is mysterious (Maykop or Yamna are often suggested but blurry).
Kristiina
Again You look superficially into modern haplogroup distribution.
Do You know the age of Serbian I2? It is called I2a-Din. It is very young. Much Younger than R1b-L23/R1a-M417 pair.
If there is a haplogroup in Europe that can be called IE without a slightest doubt it is the I2a-Din branch.
I think Balkans was a complex place. It was the place where ANE had a difficulty to infiltrate because of dense Neolithic population. We have a Thracian aDNA from 800BC he is very Sardinian shifted, so Neolithic genetics.
I believe it was the I2a_din expansion that finally overrun the remnants of non-IE population there.
The same is true for I1. But he is little bit older.
You provided that study on TMRCAs, with expansion trees. You can take a look there.
It is relatively easy to find what haplogroup expansion can fit into what Linguistic group. For example the R1b-M73 is certainly Altaic etc etc.
Dave
The original proposal saw Yamnaya directly ancestral to later cuktures like Catacomb and Poltkava . Nothing about "reflux" from areas further west (corded ware like).
It doesn't matter whether the original proposal saw Sintashta derived from Corded Ware.
What matters is that Sintashta was known to be derived from Corded Ware decades before we saw any ancient DNA from Corded Ware or Sintashta.
You can clearly see the Corded Ware "reflux" on this fairly old map.
http://s1076.photobucket.com/user/priwas/media/corded-ware1.jpg.html
I must make a correction.
I2a-Din itself could be old. It is supposed to be linked to Cucuteni-Tripolye.
But what comes under it is very young. And had a rapid expansion.
For the age of Scandinavian I1-M253 look here.
http://3.bp.blogspot.com/-FvYWmGaPfxg/VVzT6Kpe80I/AAAAAAAAKFc/BoSdLIP3mlI/s1600/table1.jpg
David
Maybe it was known, but that's not what Gimbutas, Mallory and Anthony highlight. Rather they see it as a continuation of Yamnaya by way of Abashevo . . But maybe I've misunderstood
Kristiina
This is the branch I2-L621
http://yfull.com/tree/I-L621/
I don't know exactly what they call I2a-Din.
Do they call the L621 or the I-CTS10228 I don't know exactly.
Krefter would know that better.
But You can look at the rapid expansion of I-CTS10228
It has very low TMRCA. His TMRCA is so low that it is possible to attribute it to some sort of Slavic expansion.
And in fact if You look under that branch all them are Slavs. Only one person is from Finland and one from USA of unknown European origin.
Mike, Corded Ware is unlikely to be Uralic simply because Uralic wasn't there at the right place or right time as linquistics suggests. Uralic in Bronze Age Central Europe has as much evidence going for it as Afroasiatic.
Now, if you want to criticize ADMIXTURE or other non-formal tests as measures of IE ancestry you may have a point. Saami_WGA has as much Siberian in most admixture runs as Chuvash (and tends to lack the West Asian shared between Yamnaya and Chuvash) but as per D-stats it is actually much closer to Yamnaya than Chuvash, and even closer than Norwegians, English, Belorussians etc. https://lh3.googleusercontent.com/gqnFbz-c4b2R5Fl_aKicQ-B0ZIMF5IOHh0wfKJ9uIos=w1182-h631-no
However in light of linguistic and archaeology that's trivial. Saami presence even in the periphery of Corded Ware range was barely existent (not to mention timing issues). The language has a non-IE, non-Uralic substrate but that's not likely to be from CW because other languages, all of them closer to CW range, lack it - and wouldn't that, for what it's worth, be a mark against Paleo-European CW? This was just one example but we can take any surviving non-IE language under scrutiny and the outcome - that it or its linguistic family is not CW - will be the same. More importantly, researchers of linguistics haven't found any significant evidence to suggest otherwise either.
The maps I posted are from Anthony. They show an eastern back migration from Corded Ware, resulting in Sintashta/Andronovo and eventually Indo-Iranians.
I have vague memories of someone saying that Poltavka was involved in the process, and maybe it was, but if aDNA shows that it wasn't, that's not a big deal, because Sintashta obviously was and this was known before any aDNA results.
Shaikorth
Thanks. But researchers haven't found much of any Substratum in Europe apart from criticized ideas such as the Germanic substrate hypothesis and Vasconic theories. So does this mean that Europeans were mute prior the arrival of IE's ?
But my main point was that CWC was multilingual. Maybe I erred by implicating Uralic specifically, but there's no reason to exclude it along the eastern aspects of CWC.
My further point was that CWC and BB weren't the main or final Indo-Europeanization events. IE became dominant in Europe at a much later period. Until then, it was but one of several languages, some equally far-fung, some more circumscribed.
This idea might be unnecessarily complicated, but it's closer to reality IMO
Mike, researchers found out about Saami substrate over a decade ago. But no doubt about those studies getting much less attention than something like the Germanic substrate hypothesis, especially outside academic circles.
http://www.academia.edu/4811770/An_Essay_on_Substrate_Studies_and_the_Origin_of_Saami
Thanks everyone for your inciteful critiques / comments.
It was good to bounce some admittedly different ideas . .
@Mike I agree with you.
@Alberto “Do you know if there is any study regarding this?”
I hope that Maju already responded to you. I cannot help you to find any studies on this matter. However, I agree with Maju that this influence is real.
@Aram “I2a-Din” I2a-Din is not found in Yamnaya or Sintashta, so it is only incorporated into IE’s in Corded Ware area and was probably autosomally pre-Yamnaya. The same applies to I1 although it was probably incorporated in another area and into Germanics and Finns/Saamis.
@Shaikort Saamis have probably assimilated all remaining fennoscandian hunter-gatherers + an Arctic substrate, but their core haplogroups are mtDNA’s V and U5b and yDNA I1 (in particular North Scandinavian I-L1302) and N1c (both VL29 and Karelian lines), so they are not anything exotic, but they originally probably had quite small yDNA and mtDNA pool.
For there to be replacement, there must be a continuation of mtDNA and yDNA. However, if continuation exists, it does not prove the replacement. If Eastern European hg populations were autosomally and haplogroup-wise similar, it is very difficult to prove a replacement triggered by a certain culture.
It is clear that there is no yDNA continuation between Samara R1b/Yamnaya R1b and European R1b(L51), and European IE’s are ydna-wise R1a and later on R1b, I1 and I2 + Neolithic remnants.
Samara Yamnaya (3300-2900 cal BCE) mtDNA haplogroups: H1e1a, U4a1, W6, H13a1, T2c1a2, U5a1a1, H2b, W3a1a, H6a1b
Today: H1e1a is found in Europe, North Africa, and South Central Asia. H1e1a1 haplogroup is found amongst Basques, and it is thought to be one of the autochthonous lineages to the Franco-Cantabrian Area. H1e1 is found in LBK Halberstadt and Baalberge Esperstedt. IE status in Europe: not probable.
U4a1 is found in Central Asia, Siberia and Europe. In Europe it is concentrated in Eastern and Western Slavs. However, coalescence age of U4a1 of Baltic Finno-Ugric and Volga people is clearly older than of Slavic groups. IE status in Europe: probable.
W3a1a: W3 is the most ancient W lineage and is found in Europe, including Russian regions, North Africa, Caucasus, the Near East, Mongolia and the Indian Subcontinent. IE status in Europe: probable.
W6 is most frequent in Caucasus and radiates from there to north and south. A peak of 5.3% in Georgia and the Levant, very rare in India and central Asia. Low frequency in Wester Europe <1% with the exception of Estonia and Finland. IE status in Europe: possible.
H6a1b is most frequent in Russia and the Near East and is also found in Siberia. H6 is quite rare in Western Europe and mostly found in Northern Europe. IE status in Europe: probable.
H13a1 is most frequent in Caucasus, in particular in Daghestan and in Georgia and found in low frequency in Europe. IE status in Europe: possible.
T2c1a2 is found in BA Armenia, and, today, it is mainly concentrated in West Asia. In Europe it is found at least in Southern Europe. IE status in Europe: not probable.
U5a1a1: one of the oldest lineages in Europe. IE status in Europe: not probable or at least not easy to prove.
H2b has been found in Yuzhnyy Oleni Ostrov site in Northern Karelia but it is rare in Europe today. IE status in Europe: not relevant.
Southern Yamnaya (4300-4200 BP): U5a1i, U5a1d2b, T2a1a, U4.
U5a1i: is found in ESP3 Unetice. It is found in northern Europe and Hungary, as well as in Lebanon and Sri Lanka. It is even possible that this haplogroup spread from Europe to Yamnaya. GailT suggested on Anthrogenica that U5a1i originated in Eastern European HG’s.
U5a1d2b: The "Scytho-Siberian Pazyryk culture" U5a1 samples dated 400-200 BCE appear to be U5a1d2b while U5a1d1 and U5a1d2 are most frequent in northern Europe (Ireland, Scandinavia and Russia) and we have a single U5a1d* sample in France. This probably also originated in Eastern European HG’s.
T2a1a: T2a1a is found at least in Franco-Cantabrian region and India. Instead, T2a1b is mainly distributed around the Mediterranean but also in parts of northern and eastern Europe. This is a lineage that unites Spain with Yamnaya.
U4: too widespread.
Conclusion: many Yamnaya lineages are found also in Europe. IE lineages probably include U4a1, W3a1a, W6 and H6a1b, but they cannot justify any replacement of earlier mtDNA in Europe. Southern Yamnaya is interesting as it could also justify an arrival of R1b from Europe to Southern Russia.
By contrast, Nordic Bronze Age mtDNA is very Mediterranean; in Denmark: H5a1 (H5a3 BB 10118), K1a2a (Els Trocs 10411 K1a2a, ATP3 K1a2b), K2a5, K1b1a1 (Els Trocs 10407 K1b1a1), N1a1a1a2 (5xN1a1a Starcevo, LBK), H3v, H3b (ATP17 H3, ATP 12 H3c), I, J1c4 (ATP 7 J1c1b1), T2b (3xLBK, RISE349, Hungarian MBA); in Sweden: W1 (W3 in CW and Unetice), T1a1 (RISE484 Vatya, Hungary T1a1), K1a3 (Els Trocs 10411 K1a2a, ATP3 K1a2b), J1c8a1 (ATP 7 J1c1b1), T2a1a (Yamnaya Temrta RISE547).
Swedish mtDNA is less Mediterranean than Danish mtDNA as they have W and also T2a1a is found in Yamnaya.The only Danish haplogroup that stands apart is I. Mtdna I3 has been found in Unetice, and in North Pontic steppe with a frequency of 3%.
Please, be free to complete the mtDNA picture.
Kristiina,
Modern Europeans don't cluster with Neolithic Europeans. They cluster in between them and Bronze Age steppe groups.
https://drive.google.com/file/d/0B9o3EYTdM8lQLWZnYnpwVlJrdVk/view?usp=sharing
Why do you think that is?
@Davidski “Modern Europeans don't cluster with Neolithic Europeans. They cluster in between them and Bronze Age steppe groups.”
IMO, they cluster between Neolithic Europeans and Bronze Age steppe groups because there was geneflow from East to West, but there was also geneflow from West to East as proved by Sintashta; moreover this eastern geneflow is not only from the Steppe to Europe because there was also important two-directional geneflow between the Mediterranean Basin and West Asia through the Near East.
This new paper contains a nice picture on p. 2 http://www.pnas.org/content/early/2015/09/02/1509851112.full.pdf?with-ds=yes
It shows that Sardinians, Basques, Spanish and Italians are quite close to Neolithic farmers while Scandinavians, Russians and Finnics are close to hunter gatherers and, according to your figures, also to Bronze Age steppe groups. These northern groups have higher frequencies of e.g. yDNAs R1a, N1c and Q, and mtDNAs U4a1, I, W, H6 and old hg lineages such as U5a and U5b. It looks like yDNA J1 and J2 and Q1b may have brought West Asian ANE to Europe.
@ Karl
Yes, it occured to me later that you were probably thinking of prepubescent children. Still, I think it's an unlikely explanation for lactose tolerance, because this appears to have evolved after the adoption of animal husbandry. If it were most of all about the consumption of mother's milk, then it might have evolved much earlier, during the Paleolithic. If the scenario you have in mind had any positive evolutionary value at all. But most likely the mothers wouldn't tolerate that some of their older kids disadvantage their youngest kids.
I made an error in my long comment. Instead of this.
"..Any nation who is very high in 'teal' alone is not IE , any nation who has very low 'teal' is not IE.."
There should be "some nations who are high in 'teal' ( like Caucasians ) are not IE, .."
Of course some Iranian and Indian groups are very high in 'teal' and they are IE. And this need an explanation.
Mike
I found one of Don Ringe's interview on linguistic diversity in old Europe. It was interesting and I agree with many things he say. I always show the Amerindian example for those who speak about Paleolithic languages. Thanks for that reference. But of course every idea has his own boundaries. And I don't think it is possible to apply that theory in Bronze Age period so easily.
Aram
I made specific and different points for mesolithic, neolithic /copper and bronze ages. All had different dynamics.
Lots of vivid discussions going on there, but I want to throw in some further thoughts on the fascinating ATP3 and his R1b-M269.
The intriguing thing about his yDNA is this: He has been found in a place where R1b-M269 is extremely common today. So there is something parsimonious about the assumption that there was a continuity.
And as some commenters have already suggested, a southern dispersal route for R1b-L23 is also possible, if this marker had originated in West Asia, which is possible, though still unproven.
This might explain why many Bell Beaker groups cluster with Bronze Age Cyprus in some craniometric analyses. A similar cranial series dating to the Chalcolithic was found in Paestum, Southern Italy - possibly a link.
David's PCA with ATP3 is fascinating, and it's in agreement with the PCA in the paper - it just shows more detail. Based on this, ATP3 looks like in the range of the Southern French!
I can't wait to see more analyses. This ATP3 doesn't look exactly Basque- or Sardinian-like, nor intermediate, but slightly shifted towards modern Spaniards.
@ Krefter
My mtDNA (tested by FTDNA) is K. K2 I suppose based on matches. There the HVR1/HVR2 results:
http://jpst.it/BHeU
And the known origin of this lineage is Southwestern Germany. I've got a step 0 match from Hessen-Darmstadt and one from Flanders.
@Simon_W
I agree. Many interesting opinions above, but right now we have this very intriguing ATP3 specimen that could be a really important find. I'm still pretty cautious about it, but I can't wait to see more analyses about him and the other samples (that could help to confirm or deny the validity of these initial tests).
The Spain_MN samples from Haak et al. 2015 are only some 3 centuries older and from a near by location. So this sample could represent the very initial arrival of new people. Or maybe it's all an artefact of the low quality sample. Or he's just some rara avis. But in any case, I can't wait to find out more.
One of the stats that supports the idea of these samples being shifted to the east is that until now, all EEF (including MN) had more shared drift with Sardinians than with any other modern population (followed by Basques). The f3 stats with outgroup from Haak et al. confirm this (Figures S7.5 and S7.6). However, the ATP samples, 3 out of 4 tested, share more drift with Basques than with Sardinians in the same type of stats (Figure 3B). It can't be because of the higher WHG, because the Haak et al. MN samples already had as high or higher WHG ancestry than these ones. So it could be related to ANE.
@ Alberto, interesting observation. Another factor lowering f3 might be new Neolithic ancestry from a different source (e.g. whatever tended to come in with the EHG at the same time).
For example, in Figure S7.5 from Haak et al, Loschbour has a higher f3(Dinka,Pop1,LBK) closer to LBK than any modern North European. Also true for the f3 stat with all the Copper Age and MN populations (even Motala has a higher f3 with Spain_MN than Sardinian!).
Notably the West Asians do not even enter the picture with these stats - if I understand correctly, that's because most of these Middle East / Neolithic populations are quite genetically diverse and so do not share a lot of drift with one another.
Following may be tl:dr, but with outgroup f3s, I think the way it works is outgroup f3 stats of the form f3(outgroup,Pop1,Pop2) are basically a measure which say "At those locations where Pop1 and Pop2 both differ from outgroup, how often are they similar?". This is different from the raw IBS between Pop1 and Pop2 in terms of IBS just says "How identical by state (same frequency of variants) are these two populations?" without using the outgroup to "control" for differences in drift between 1 and 2 (note that this does not control for drift between say, a hypothetical Pop3 and Pop2 or Pop3 and Pop1 at the same time, just between the 1 and 2 that are in the stat).
The way that the Reich group put it is, paraphrasing http://genetics.med.harvard.edu/reich/Reich_Lab/Welcome_files/2012_Patterson_AncientAdmixture_Genetics.pdf - page "1067", for an f3 stat f3(C;A,B), the stat is the sum of the difference in allele frequency C-A multiplied by C-B. When one population has 0 difference from C, then C-A*C-B will be 0 as well (0 multiplied by anything is 0). So the more often this averages out to no difference due to low levels of drift, the lower the f3 stat, while the more often the difference C-A is in the same direction to C-B (same +/- sign), the higher the stat will become.
This does mean that the stats f3(outgroup,Neolithic,Neolithic) for instance will tend to be lower than f3(outgroup,HG,HG) or even f3(outgroup,Neolithic,HG), simply because Neolithic populations seem to retain a lot more of the ancestral (African) diversity. Either through less drift or more admixture offsetting drift. So they either have lower diferences from Africans and where they different from Africans are also different from one another = lower f3.
I think if I'm understanding correctly, you can see that in Haak's Table S7.1. The Neolithic Europeans have much less maximum f3 with one another than the HGs, even though by the standards of many modern West Eurasian populations with one another, they may be share lots of f3 drift. I haven't seen any systematic comparison of these stats to compare, so that's an assumption. (There's a difference in the HGs though, as well, where Samara and Karelia share less with one another than say Losch and La Brana, and more like what is shared between Losch and Motala).
Also there's that pattern in the Figure S7.7 LNBA f3 stats where LBK_EN and Spain_EN generally have more shared drift with Bell Beaker, Halberstadt and Unetice than many modern North Europeans, while Sardinians are nowhere to be seen (Basque are on the list though).
@mike
the kurgan model has already failed
I don't know which particular model you are arguing against. Most of these things have not been said by David Anthony who is my main source. If you are referencing Gumbitas' then I will tell you upfront that I think she is crazy and sexist as hell against men, hence I don't think she is reliable.
As far as I know David's kurgan model is perfectly compatible with a southern influence in the yamnaya. Its something I expected. If yamnaya is late PIE, and PIE is Caucasian influenced, it makes sense that yamnaya are themselves have southern admixture.
David does not describe a direct chain of "Yamnaya -> Catacomb -> Andronovo". I have the impression he would support "CWC- >Sintashta -> central Asia".
In a nutshell he predicts Yamnaya (or their predecessors) > Afanaseivo. Yamnaya > Central/Western Europe. MINOR Yamnaya input into CW.
Now, David talks about migration and the "direction" of cultural influences but he was not talking about genetics directly. However, the chains he gives are panning out genetically (even though they didn't really have to). Thus far, the DNA results are pretty much exactly as I had predicted and I have been using davids model.
David has been a stupendous resource for predicting genetics. This cannot be overlooked. He said the yamnaya migrated out of the steppe and to the Altai in one direction and to central europe in the other direction. Ancient dna tells us that a yamnaya like people migrated to the altai and central europe (bell beakers have the autosomal and ydna relation to yamnaya folk).
Ancient DNA has not yet proved it was THE yamnaya doing these migrations but I'm already in AWE. Already the yamnaya are close in time and space to the actual migration that we will soon confirm.
David seems to be correct about the chains that you mention but because I'm mostly only concerned about the genetics I would have forgiven him made some erroneous links. I only care that he got the players right and the general timing down. You care a whole lot about the mechanisms of social change so I can see how you would care about the specific chains being used. I think this different emphasis is why there has been some misunderstandings. For this hobby I'm mostly a where and when kind of guy :)
@david,
are there any updates on how certain we are that ATP is r1b, much less m269, and how certain we are that he has yamnaya like influence. Chad said he would do an estimate but I haven't seen any further results.
We need to figure out why the study did not publish ATP's ydna
@Colin
They didn't publish autosomal DNA or Y DNA of any of the 4 poor quality samples, because they were poor quality.
Colin
You're correct I was mostly critiquing Gimbutas. Anthony's models are far more sophisticated. Nevertheless, for a few issues both the 'general gist " and the specifics need amendment. Now, that's not a problem. That's science & progress . .
As Karl was saying earlier, my different opinion is a matter of degree.
Or precision (Id like to think).
But enough for now :)
Kristiina said: "I hope that Maju already responded to you".
Well, he was not asking me (just checked) but in any case his question was about Basque phonetic influence in Castilian (alias "Spanish"). My knowledge come from primary education, where in textbooks (Spanish textbooks) it was considered a matter of fact by Spanish linguists of the time (1970s). Some things are obvious anyhow: loss of the sound /v/ in favor of /b/ (Basque does not have the sound), loss of initial f- (what other Iberian Romances retain, /f/ is not a Basque consonant even if today is found in few anomalous words) and the 5-vowel system, which, although it did exist apparently in Latin (if we ignore the short/long vowel difference only), is absent in most other Romances, which have more than just 5 vowels. This 5-vowel trait could be the reason why some ancient author, whose name I don't recall right now, praised pre-conquest Aquitanians for their excellent Latin, much better sounding than the one of Rome back in his day, which was surely already evolving to Vulgar Latin. Some of these traits are also present in Gascon (which is not really "Occitan" but a distinctive Romance).
In addition to that, Ibero-Romances have two distinct "to be" verbs (one for existence or having intrinsic qualities: "ser", another for the locative or transitional qualities function: "estar"), while Gallo-Romances (nor Latin) do not. However the widespread legacy of this trait in all Ibero-Romances, including the transitional Catalan/Occitan, rather precludes it from being a Basque-specific influence and should be considered generic Ibero-Vasconic instead.
"Modern groups speak many different languages: Armenian, Northeast Caucasian (Avar, Tabasaran), Turkish, Bashkir Turkish, Arab, Chinese, Urdu/Hindi, Kurdish, Slavic and Romanic languages. Georgian should probably also be in the list. I do not know if the case is very strong for IE identity of Yamnaya. I would put my money on a NortIheast Caucasian identity."
?????????????????
About half the languages you mention are IE (Armenian, Hindi/Urdu, Kurdish, Slavic, Romance) and what large group of R1b's speak Chinese except maybe Uyghirs of former East Turkestan forced to learn Putonghua by a colonial government? Turkish of course spread in Medieval times so would not have been spoken among populations in regions like Turkey or Bashkirostan prior to the middle ages, Arabic likewise is known to have spread far in the middle ages. I wouldn't be surprised if some R1b were Caucasian native speakers but the link to IE seems much clearer than to Caucasian.
Hopefully you're saving any Ancient matches among fully-sequenced Ancient mtDNA.
Anyways, I found 100s maybe 1000s of fully-sequenced Danish mtDNA. I already have low-coverage mtDNA from most of West Eurasia to give a broad picture of mtDNA diversity. Fully-sequenced mtDNA like this will reveal a lot of substructure for all of Europe not just Denmark.
http://www.ianlogan.co.uk/checker/all-li-sequences.htm
You can find their rCRS mutations here.
http://www.ianlogan.co.uk/checker/checker-li.htm
Ian Logan has tons of fully-sequenced mtDNA you should look into it. Looking at Danish H1 I can already see a lot of differences with Basque H1(I have 426 full-sequenced Basque H).
Could be because of a separation of H1 lineages that occurred in the Early Neolithic Balkans: One set of H1 migrated via Mediterranean to Iberia and one set migrated via mainland Europe to Scandinavia. Funnel Beaker Gok2 belonged to H1c which is rare in Basque and popular in Danish.
@a Res You IEs are very possessive about yDNA R!
However, I repeat my logic. R1b-269(xL23) is found in Turkey, Belarus and Sardinia. It is very ancient (formed 13300 ybp, TMRCA 6400 ybp) and precedes any expansion of IE languages. The fact that it is found in Turkey and Sardinia favours its origin in Neolithic Turkey/Eastern Mediterranean. This line is also found in RISE524, Mezhovskaya, and nobody has claimed Mezhovskaya to be IE. The next level is R1b-L23(xL51) which has four branches: 1. Turkish/Moroccan Jews; 2. Armenian, Assyrian, Sardinian, Sicilian and RISE397, Late Bronze Age Armenia, Kapan; 3. Avar branch that is found in RISE550, Yamnaya, Peshany 4. Tabassaran branch, Pathans, Indian, Chinese, Burzyan Bashkirs, Kurd +European branch, and RISE547, Yamnaya, Temrta IV and RISE555, Stalingrad Quarry. Nr 1 is non-IE, Nr 2 is probably non-IE as it is only found in Armenians and other ethnic groups are non-IE or only recently IE (Sardinians, Sicilians), nr 3 is specifically North-East Caucasian, Nr 4 is highly international and European R1b is only a small sub-branch that looks like having originated in Bulgaria that has never been assumed any centre of IE expansion.
When I wrote that mtDNA I has been found in Pontic Steppe, it seems to be a different subclade from I3.
Yamnaya (3600—2300 BC)
SUG2 - I1a, Kirovograd Sugokleya Central Ukraine
I1a has been found in Unetice.
Catacomb Culture (2800—2000 BC)
TET1 : I1d, Tețcani, Moldova
Catacomb mtDNA:
U? 1/24, 4%
U4 7/24 29%
U5a1 4/24 16.6%
H (rCRS) 1/24, 4%
H2 1/24 4%
H1, H3 or H6 3/24 12.5%
H1 or H13 1/24 4%
H6 1/24, 4%
H (in total) 28%
J2b 1/24 4%
J1b1a1 2/24 13.3%
R1a 2/24, 8%
I1d 1/24, 4%
They have a terribly high percentage of U4 and no T. As Ukrainian and South Russian Yamnaya both have T, it is probably a coincidence. Hope we will soon get their yDNA.
Kristiina
Didn't the paper by Der Sarkissian / Haak suggest that catacomb appears to be a genetically more "forest" looking culture, even some kind of replacement of Yamnaya. Of course, genome-wide data will be the clincher
@Kristiina: very well said!
@Krefter: I hope to see soon a synthesis "for dummies" (not too dummy, you know) of your work on mtDNA: looks most interesting.
@ Mike Thomas - "catacomb appears to be a genetically more "forest" looking culture"
I think that's mtdna pattern, which could be quite different from the y-dna or whole autosome.
To take a probably superficial read, these early steppe / kurgan / Indo-European cultures seem like son obsessed cultures, judging by the presumable spread of their patrilineages. (Like cultures of Henry VIIIs, only without so much of his bad reproductive luck). It seems they may have been very driven to ensure that fathers have sons, and those sons have sons of their own, etc. perhaps to the extreme neglect of devoting any resources to the survival and reproductive success of daughters, and ultimately the continuation of their maternal lines.
If you persistently devote all your family resources to your sons to make sure they have sons, your y-dna haplos will in time replace everyone who doesn't go hard with this "strategy" (or culture, really). They may even cross language and cultural barriers to make this happen (if your cultural drive is so strong that sons will even switch language and eventually lose their culture's mythos to maintain their tradition of patrilineage).
But all things being equal, your mtdna will probably disappear, to be replaced by those from foreign women. (While the autosome will change to be a mix). All things being equal, if you don't have any especially high ability to gain resources overall.
"Old European" (to use probably a lumping category) farmers and the HGs may not have been as much like this - if not matriarchal (in the sense of pacified or subordinate men) or matrilineal, they may have at least had some greater concern with in group female fertility and having daughters and keeping the lines of the ancient mothers going (possibly overgeneralising from those mother goddesses etc.) Although possibly the "Old Europeans" may have gone through cultural cycles like this, to explain some of their turnover in mtdna vs ydna.
This wouldn't necessarily have to be a total difference, just one with a lot of compound potential over time or under conditions of cultural / environmental stress.
@Matt
Thanks for the detailed explanation.
With these ATP samples it's still difficult to know what to think. I wonder about how the quality can be affecting the few tests we've seen. But if we accept the results, the slight change in the f3 stats would still suggest some new population in the mix (Armenian-like?). As Simon said, it's intriguing because of how common is R1b-M269 nowadays in Spain, because in the PCA the samples does cluster between Spaniards and Basques, and because we have the traditional view of Bell Beakers originating in Iberia. Of course it's still quite difficult to explain how this population could arrive to Spain without leaving traces in Hungary or North Italy, where we have younger samples that still look like European Neolithic farmers. But in any case, the results are interesting enough to investigate a bit more about them. I hope David has time to post further tests with all the samples so we can see if there's something real about it or not.
Matt
Yes I agree. That's why I said we need full genome analysis. And I agree with you about fathering & patrilocality- i had thought this was an overstated cliche. But I really think now it played a role. It might also explain why Yamnaya and corded ware like groups expanded: local tensions over resources and rule of ones household must have required an "exportation of conflict"- ie migration.
Alberto/ Dave/ Matt
Didn't some earlier looks by David suggest EHG like admixture -albeit minor in older samples like vinca and even Loschbour? I think we might have dismissed them as noise . Maybe worth looking again at?
ATP3 is difficult to make sense of because it only offers around 3K transversion SNPs.
I'm trying to get ATP2 now. It's a much higher quality sample, at around x4 coverage, so it should have well over 50K transversion SNPs available.
It's a bit younger than ATP3, so if there was eastern gene flow into northern Spain during the Copper Age, it should still show some.
@Maju
Thanks for all the comments. Yes, indeed I was referring quite specifically to the phonetic influence more than to similar words, toponyms, etc... so your second comment was spot on.
What intrigues me about this is the fact that this "Basque-like" phonetics are somehow common in Southern Europe. I mean, Latin itself already has rather similar phonetics. Italian is not very different either.
I think that phonetics are an interesting part in language transmission. For an easy to understand example, we can simply compare how when the Germanic speaking Franks switched their language to Latin, they retained a good amount of the Germanic phonetics, which is why French sounds clearly different from Spanish and Italian (and much less Basque-like, if you allow me to use this unscientific expression).
I don't know how familiar you are with other Southern European languages, but if you've heard enough modern Greek (or at least studied some classical Greek), you might have also can't help to notice that the phonetics don't differ that much from Basque either (especially when compared to, say, Swedish or Danish). Also quite interestingly, southern Slavic languages (Serbian, Bulgarian, and most of all Macedonian), also share some of these traits. Macedonian more often than not sounds (phonetically) closer to Spanish than to Russian or Polish, even if the language is obviously much closer to the latter ones.
So I was wondering if there was any study about this seemingly common phonetic substrate in Southern Europe. I never heard about any, and apparently Kristiina neither, but I think it's a subject that could yield some interesting results.
@Kristiina
However, I repeat my logic. R1b-269(xL23) is found in Turkey, Belarus and Sardinia. It is very ancient (formed 13300 ybp, TMRCA 6400 ybp) and precedes any expansion of IE languages. The fact that it is found in Turkey and Sardinia favours its origin in Neolithic Turkey/Eastern Mediterranean. This line is also found in RISE524, Mezhovskaya, and nobody has claimed Mezhovskaya to be IE. The next level is R1b-L23(xL51) which has four branches: 1. Turkish/Moroccan Jews; 2. Armenian, Assyrian, Sardinian, Sicilian and RISE397, Late Bronze Age Armenia, Kapan; 3. Avar branch that is found in RISE550, Yamnaya, Peshany 4. Tabassaran branch, Pathans, Indian, Chinese, Burzyan Bashkirs, Kurd +European branch, and RISE547, Yamnaya, Temrta IV and RISE555, Stalingrad Quarry. Nr 1 is non-IE, Nr 2 is probably non-IE as it is only found in Armenians and other ethnic groups are non-IE or only recently IE (Sardinians, Sicilians), nr 3 is specifically North-East Caucasian, Nr 4 is highly international and European R1b is only a small sub-branch that looks like having originated in Bulgaria that has never been assumed any centre of IE expansion.
All of these R1b cases can be explained by invoking IE origins given the genetics and archaeology/history of these populations and regions. I am not saying all of these R1b are necessarily from IE, but they might very well be.
Alberto
Not about Europe and perhaps not what You search but related Global research.
http://www.pnas.org/content/112/5/1265.abstract?sid=926ec0d1-136c-4035-94cd-47726d2d0d3e
and this map
http://www.pnas.org/content/112/5/1265/F1.expansion.html
@Alberto
I don't know how familiar you are with other Southern European languages, but if you've heard enough modern Greek (or at least studied some classical Greek), you might have also can't help to notice that the phonetics don't differ that much from Basque either (especially when compared to, say, Swedish or Danish). Also quite interestingly, southern Slavic languages (Serbian, Bulgarian, and most of all Macedonian), also share some of these traits. Macedonian more often than not sounds (phonetically) closer to Spanish than to Russian or Polish, even if the language is obviously much closer to the latter ones.
You can add Hittite and similar Anatolian IE languages too to the list of IE languages with a Basque-like phonetics. Hattic (the pre-Hittite non-IE language of central Anatolia) probably had a Basque-like phonetics as well.
@Alberto: You're welcome. The issue of coincidence of the 5-vowel system in Latin and Basque is somewhat intriguing but I can't judge it properly. Vowel systems vary a lot: from a total of 20 cataloged possible vowel sounds (IPA), some languages have as few as two, while others like Danish have as many as 15. Single linguistic traits as these, taken alone, are usually not too informative, however when added up to other traits they make constitute a pattern.
What you say about (modern) Greek is interesting indeed. However we should also consider here the Etruscan/Pelasgian connection (which surely influenced Greek and Latin from a common source) and this presumptive family does not seem to have any direct connection with Basque or Vasconic (although it surely has very old Vasconic substrate, if my theory is correct). Ancient Iberian for all we know should also have a 5-vowel system, much unlike the most usual Catalan dialects of today. But not much to say about it because it's just a trait, and we cannot conclude much based only on it, just raise an eyebrow at the "coincidence".
A possibility could be that areas most affected by so-to-say "strong Indoeuropeanization", which would be rather Northern Europe, tended to lose more of the traits that were once more common in (Neolithic) European languages. The pattern is however irregular, as we can see in Romances. Why does Portuguese have so many vowels? Celtic influence, more recent French one? Why Italian does instead keep a more constrained vowel range? Is it a peculiarity of Tuscan or did it affect all dialects before unification? Is it early Modern Spanish influence even? Hard to say. At least I don't have a qualified opinion, just open questions.
Yes, it would be interesting to see Catacomb yDNA; if it is not again R1b. The high frequency of U4 is a northern trait. Their I1d appears to be very a very rare lineage. I find J2b interesting as it has been found in CW, Unetice and Sintashta, Mezhovskava and Salzmünde, but J2b1 appears to be “virtually absent in Europe. Found in diverse forms in the Near East” (Wikipedia). I could not find a better link to the distribution of J2 but here you go http://www.joanannlansberry.com/other/bol_dna.html
There is a big concentration of J2b in Ob-Ugrics who have been connected with Mezhovskava Culture: Mansi J2b 6.2%, in another study even 10%, U4 15.5%, Khanty J2b 3.8%, U4 8.5%, Ket J2b 0%, 30% U4, Selkups J 0%, U4 40%, Nganasan J 0%, U4 21%. Also Kalash have 10% J2b.
Onur, of course it is possible, but I think that all existing language families have been succesful and wide spread at some point of time. Today Celtic languages are mainly spoken by a few villagers in remote areas in Ireland and Great Britain, but they were much more wide spread and powerful before. The same goes for Ob-Ugric languages. Today, they are spoken only by a few remote groups on the River Ob but before their had their duchies in Yugra on Pechora River. In the same way, I think that Caucasian languages were much more wide spread before, but have receded and their area has shrinked to small patches in the vicinity of the Caucasus mountains. IMO, we underestimate the amount of different languages and extinct languages in the areas that are today dominated by recent expansive language families, such as IE languages and Turkic languages. I think that before the modern era language maps were like the linguistic map of northern NA languages: https://en.wikipedia.org/wiki/Indigenous_languages_of_the_Americas#/media/File:Langs_N.Amer.png
@David
Yes, I think that looking at ATP2 would be the best way to see if the ATP3 results are real or just due to the low quality sample.
@Mike
If i remember correctly, the Vinca sample showed that EHG-like admixture only in a test that David made that was not good for samples with high WHG ancestry (any excess of WHG ancestry over Germnay_MN showed as steppe ancestry, even in Loschbour as you mention). But that was just a technical limitation of the test. Still, that Vinca sample would be worth a second look, but I think it has the same problem as this ATP3: low quality, which makes the results not too reliable.
@Kristiina: North America is 2.5 times the size of Europe and we can safely assume (IMO) that the Western and Central regions were almost uniformly speaking a single language family (IMO Vasconic), excepted some areas of the Balcans (Vinca-Dimini's "Pelasgic", always IMO). So it only leaves the Eastern third really open for speculation and there we can pinpoint two mayor language families (IE and Uralic), and some minor ones in the Caucasus presumably. The main open question would be the language of the Dniepr-Don culture, with apparent roots in local Epi-Gravettian, and the probably related Pitted Ware groups of the Baltic.
So I would think that, excluding North Caucasus and possible pockets of HG persistence in the West, we would be talking of five families (or so), three of which have modern living descendants, another might have a historical descendant in Tyrsenian (Etruscan, Lemnian) and the last one would be the Dniepr-Don + Pitted Ware one, of which we know nothing.
It's a model that could be a bit simplistic but at least as a draft on which you can apply your own ideas, it should work.
Notice that not just North America is much larger than Europe but also there many populations were barely Neolithic at all (no herding, few crops available) and the North American Neolithic spread is a clear case of diffusion, while European Neolithic is now a clear migrationist scenario instead, with only the Eastern European partial exception to fall maybe to the diffusionist side.
Even with all that many areas of North America had a rather simple linguistic organization, the main exception being the mountainous Western Coast. Some of the family distinctions are also surely caused for lack of data rather than strict demonstration of non-relationship.
@Aram, Onur
Thanks for the link and input. I'll take a look at those as soon as I have some time.
@Maju
Yes, the case of Portuguese is a bit strange, since the phonetics are quite different from the pattern I'm referring to. I really don't know where did Portuguese get some of these phonetics from, which are quite alien to Latin, Spanish or Basque.
It's a pity that I don't remember well much of what I studied at university (yes, I did study philology, though I didn't finish it and I'm in no way a linguist), because then I could point out with more precise terms the patterns I refer to. The 5 vowel system is one of them, yes, but the pattern extends other features that make these languages "hard" sounding (cf. with North European ones). It's quite obvious by just listening to the languages (for example, I don't understand a word of either Basque or Greek, but both sound perfectly easy for me to pronounce. I hardly speak Macedonian either, but when I do have to speak something, people say that I sound like a Macedonian, while pronouncing something in Russian -not that I speak it either- is quite more unnatural for me).
I never thought much about this before, but suddenly your comment about seeing Basque influences all around did ring this bell in my mind. Specially because I realised that it applies quite clearly to the south of Europe, no matter the language family people speak.
I'll see if I can find anything more about it.
@Alberto: you probably won't find much because these subjects are very obscure, Basque linguistic academy is largely entrenched in the myth of uniqueness (even rejecting Vasco-Iberian that only recently has made a much needed comeback), the dominant prehistoric paradigm was until recently one of "paleolithic continuity" (barring IE expansion) and Neolithic diffusionism (at least in the South, less defensible for Central Europe) and non-Basque linguists only rarely are interested in Basque, let alone tracking Vasconic substrate. The result? The less bad theory we have is Venneman's Vasconic substrate theory but this one is trapped by design in the hypothesis of Paleolithic continuity. So it's useful but also limited, particularly re. the South, which he mostly considers (capriciously IMO) outside the Vasconic area.
@Kristiina
I think Native American levels of language family diversity existed in Eurasia and Africa only during the hunter-gatherer times. Beginning from the Neolithic migrations, language family diversity has largely been razed in Eurasia and Africa, so thousands of years before the modern era there was already much less language family diversity in Eurasia and Africa than in the Americas.
Onur
Language diversity would have risen after the Mesolithic, and peaked during the Late Neolithic. It only began to reduce in the Bronze Age (>2300 BC), and more rapidly after the start of the Iron Age .
@Mike: so, according to you, all those genetically homogeneous EEFs dedicated their sparse free time between clearing the lands, plowing them, harvesting the grain and processing it, among many other labors, to linguistic engineering so their once common language diverged much much faster than is usual? Possibly you imagine that they managed to achieve such linguistic evolutionary speeds that in a mere thousand years they evolved like normal languages do in 10,000.
Else what? Just making a claim is meaningless, and your lack of arguments does not say anything in support of your opinions. Not now, not in many other instances.
I say that the genetic homogeneity that EEFs (and largely their partial descendants the "MN" type populations) is not only evidence of a migrationist paradigm but that such a migrationist scenario necessarily implies the expansion of a single linguistic family rooted in Thessaly's Neolithic (or a wider Aegean one if you prefer). Even the partial second HG admixture we see towards the Atlantic is surely not enough to grant survival to pre-Neolithic languages, because the immigrant fraction (quite large) would always have the advantage of "cultural superiority". Only the remnant bands of HGs would retain pre-Neolithic languages, maybe not even in some cases, and their destiny was anyhow extinction or assimilation.
We can imagine many shades and twists on top of this generic draft but hardly an increase of linguistic diversity in all the area of mainline Neolithic, which included most of Europe at the beginning of the Chalcolithic, rather the opposite: a single linguistic family expanding much faster than it had time to diverge. IMO when the La Hoguette and LBK farmers met at the Rhine, some 1000 years after they parted ways in Greece, they could still understand each other almost certainly, although with some difficulty.
Language diversity would have risen after the Mesolithic, and peaked during the Late Neolithic. It only began to reduce in the Bronze Age (>2300 BC), and more rapidly after the start of the Iron Age .
Can you give references for these claims Mike?
Basque vs Dansih
http://mtdnaatlas.blogspot.com/2015/09/basque-vs-danish.html
At the basal level Basque/Danish H is basically the same, they fall under the same European cluster(as opposed to West Asian H which is pretty differnt). But within European-clades(like H1/H3) they're very differnt. It looks like 50%+ of Basque/Danish H split from each other before or during the Neolithic.
I'm not 100% for any particular theory on the origins of H in Europe. It's incredibly diverse and can have a million differnt origins. A Neolithic origin for most makes the most sense to me right now, but there's plenty of room for Pre-Neolithic H. We have prove H did exist in Pre-Neolithic Europeans who lacked Near Eastern ancestry.
There isn't enough pre-Neolithic mtDNA to prove the vast majority came in the Neolithic. If the loads of H from pre-Neolithic Iberia are confirmed that'll be huge. There may have been H-rich WHG groups who were absorbed by EEF or rare WHG-H lineages that went through founder effects and became very popular today.
Seems like the Caucasus is the best null model for pre-IE expansion European language diversity in Europe? Probably migration based founding of the farming communities there. Differences would emerge in the mountainous structure there, affecting differential diversity. But that might be offset by how much harder travel was in an earlier era.
I don't imagine anyone can really estimate pre-IE European language diversity, though. Where is any real evidence?
It is probably true that a big language family (families?) formed in Europe with Neolithic migrations and everywhere with Bronze Age kingdoms. However, I did not propose that Southern Russian Yamnaya spoke a hunter-gatherer language but North-East Caucasian languages that have a well developed pastoralist lexicon (e.g. words for animals of different age, heifer, ram, manger, cattle shed, watering trough, mountain pasture, etc.).
Maju, that was a good summary. However, Caucasus is full of entire language families that IMO should have been more wide spread before. On the northern side, we have North East Caucasian language family and the North West Caucasian language family which is structurally clearly very different from the first one. On the southern side we have Georgian and other Kartvelian languages. Historically we know that there existed Urartuan and Hurrian languages (linked with North West Caucasian language family) and the Sumerian language spoken south of Caucasus Mountains. It has been proposed that proto-IE has a Caucasian substrate, so it would not be so bad idea to propose that Caucasian language was spoken in Southern Yamna (Rostov and Volgograd oblasts)
To me, it looks like IE was introduced relatively late to the area. Sumerian Empire 3350 BC, Hurrian empire c. 2300 BC, Hattian Empire 2300-1600 BC, Hittite Empire 1600-1200 BC, Mitanni Empire 1500–1300 BC, Urartu 800-500 BC (according to Wikipedia Urartuan was replaced by Armenian language after that). Hittites are IE and Mitanni had an Indo-Aryan ruling class, but around Van a non-IE language was still spoken 800-500 BC. Southern Yamnaya samples are dated c. 2300 BC, so it is contemporaneous with the Hurrian empire and precedes Hittites 700 years. Davidski has proposed Repin Culture and River Dneper as the area of origin for R1a and that would fit quite nicely with Southern Yamna being Caucasian. However, I admit that we need more yDNA to be more certain about distribution of languages.
Historically we know that there existed Urartuan and Hurrian languages (linked with North West Caucasian language family) and the Sumerian language spoken south of Caucasus Mountains.
They are actually linked with the Northeast Caucasian language family not with the Northwest Caucasian family, but I have not seen a strong case for the Sumerian link.
Onur
Yes; I've already mentioned them above; but see in particular 'A Social prehistory of IndoEuropean..' By John Robb.
Maju
Yep the genetically similar early farmers were lingusitically heterogeneous . That's exactly what I'm saying, and that is what ethnographic evidence lends credence to; just see above ref. I know you won't like that theory because it means that Basque wasn't in fact spoken everywhere from Britain to the Caucasus back in the gold ole' Neolithic golden era.
In it didn't require "linguistic engineering" , just natural history.
Onur, that is interesting! North-West Caucasian groups have clearly much less R1b. They have mostly J and G and R1a (mostly Indo-Aryan Z93).
I googled that some North-East Caucasian groups have high frequencies of R1b: Bagvalins 68%, Tabassarans 40%, Lezgins in Axtynskiy District 30%, and, as I have already noted, in R1b-Z2103 tree, there are even two lines that are specific to them, Z2106* Tabassarans, and 15080170 (C/T) and 18945942 (C/T) Avars. Armenian R1b average is 30%.
I think that the Caucasus diversity reflects a (partly documented) greater diversity of language families in Neolithic West Asia because there the diffusionist model applies much better and there is only limited evidence of single-origin expansions before the Semitic and IE ones (the latter with three apparent distinct origins: north for Anatolian, east/NE for Indo-Iranian and west for Phrygio-Armenian quite probably).
The linguistic landscape of Neolithic West Asia was hence probably more diverse than the European one, much more similar to your Native American example. But owing to founder effects, the impact of this diversity in the periphery of West Asia, namely Europe and India, was surely one of large homogeneous language families (mostly): probably Vasconic in Europe and Dravidian in South Asia, this one may be related to Elamite.
"It has been proposed that proto-IE has a Caucasian substrate, so it would not be so bad idea to propose that Caucasian language was spoken in Southern Yamna (Rostov and Volgograd oblasts)".
My own modest private research rather suggests that PIE has weaker relations to NE Caucasian than to Basque (these owed possibly to shared Paleouropean substratum), at least vocabulary-wise, but I never got to compare with Kartvelian or NW Caucasian. In any case the contact zone could well be the North Caucasus itself, where Khvalynsk culture also existed, producing both Yamna (Volga) and Maykop (North Caucasus) as local offshoots. I hate to but I must insist on Yamna not being the very beginning of Indoeuropean but rather a second or third stage in the same urheimat area. It seems quite plausible to me that Maykop and its close relative Kura-Araxes cultures not only included IE (Anatolian branch) but also Caucasian speakers, that it was probably a multilingual culture that never really got homogenized in this aspect (although it probably had a lingua franca). It's possible that something similar happened in other places in the first phases of IE expansion, at least where demic replacement was not significant, but it's harder to know.
Mike
Reconstructed PIE *gu/gwau- ‘cow, ox’.
Sumerian cuneiform first attestation 3000 BC - gud/gu ‘bull, ox, cattle
Modern English - cow, modern Armenian kov
Amazing stability (the slight meaning shift not counted) for a word who is aged 5000 years non?
But taking into account that PIE was not in Mesopotamia this word needs some extra time to travel from elsewhere. This will easily add some 1000 years. We have 6000 years at minimum.
I don't think it is possible to compare the situation in Pre-Columbian America to Eurasia. Some will say Steppe was different. But it is possible to find words as far as in Altaic Turkish without much phonetic change from the attested cuneiform version.
@Kristiina
I googled that some North-East Caucasian groups have high frequencies of R1b: Bagvalins 68%, Tabassarans 40%, Lezgins in Axtynskiy District 30%, and, as I have already noted, in R1b-Z2103 tree, there are even two lines that are specific to them, Z2106* Tabassarans, and 15080170 (C/T) and 18945942 (C/T) Avars. Armenian R1b average is 30%.
Only southernmost Northeast Caucasian speakers (the Lezgic branch), who have been neighbors of IE speakers or people significantly descended from IE speakers for thousands of years, have a high percentage of R1b, the rest of Northeast Caucasian speakers have R1b in pretty low levels.
Kristiina
You forgeted to mention the levels of J1 and J2 (Nakh branch) that North East Caucasian people have. Where is the J1 in Yamna? 0%
Trying to link R1bwith a North East Caucasian languages is a lost case. If You don't believe me just reread the Balanovski et al. recent update. The Caucasus is a very nice place to figure out what Y DNA was connected to what linguistic family.
http://mbe.oxfordjournals.org/content/28/10/2905
If You are uncomfortable with linking R1b-L23 with IE call it something else. For example Unknown Language Family (ULF) like the UHG.
Or You can call it Vasconic. What You think Maju?
BTW If You can read Russian I can send You a link where You will learn that Burzyan Bashkirs are also called Alan-Burjans because they claim to have an Alanian origin, and it is recognized by other Bashkir clans.
The amazing thing that there is another nation Osetians who claim an Alanian origin. And guess what they share the common R1b Z2103+ Z2109+ subclade. :)
The other thing with Ossetians is the fact that they don't have practicaly the Iranian Z93. The only clade that links them to the IE world is the Z2103+. The rest is the Caucasian G2a.
It's interesting that in the ADMIXTURE analysis of the paper, at K = 6, there appears a Kalash-related or rather generally ANI-like pink component which is present throughout all IE populations of Europe, but only patchy in Basques and almost absent from Sardinians. In the ATP samples this component is only present in the Middle Bronze Age female ATP9. In her it is more present than in modern Basques. So I think we can safely say that this pink component is related with the incoming IEs, but not with the first Basques. The fact that it's absent from all earlier ATP samples suggests that these were not IE, but may have been Basque-related.
@Simon_W
I don't quite like that ADMIXTURE run in the paper. A good amount of the ANE in Europeans seem to be in the blue component, so in populations with low amounts of it it gets hidden. David's latest K12 didn't suffer from this problem, but it seems ATP3 has just too few markers to run it through it.
In any case, the paper's Admixture does show is that we're talking about low levels of ANE (if any), most likely in the Basque range. Now I checked the IBS from Spain_MN from Haak et al. and it's the same pattern as with the f3 stats with outgroup: Spain_MN (like all other Neollithinc European farmers) have Sardinians at the top. Here for the 3 Spain_MN samples:
Spain_MN I0406:
Sardinian 0.735087
Spanish_Valencia 0.734108
Spanish_Aragon 0.733647
Basque_Spanish 0.733639
Basque_French 0.733525
Spanish_North 0.733157
Spanish_Castilla_la_Mancha 0.733080
Spanish_Cantabria 0.732924
Bergamo 0.732796
Spanish_Extremadura 0.732568
Spain_MN I0407:
Sardinian 0.677759
Basque_Spanish 0.676900
Spanish_North 0.676478
Spanish_Cantabria 0.676157
Basque_French 0.675816
Spanish_Aragon 0.675072
French_South 0.675006
Spanish_Valencia 0.674398
Spanish_Extremadura 0.674265
Spanish_Castilla_la_Mancha 0.674208
Spain_MN I0408:
Sardinian 0.738277
Basque_Spanish 0.736405
Spanish_Valencia 0.736137
Spanish_North 0.735872
Spanish_Castilla_la_Mancha 0.735840
Spanish_Aragon 0.735683
Bergamo 0.735679
Spanish_Cantabria 0.735275
Basque_French 0.735254
French_South 0.735070
In contrast, APT3 results posted by David:
Basque_Spanish 0.830285
Sardinian 0.829974
Basque_French 0.829737
French_South 0.829656
Spanish_Murcia 0.829611
Spanish_Cantabria 0.829379
Spanish_Cataluna 0.829140
English_Cornwall 0.829034
Croatian 0.828996
Spanish_Valencia 0.828771
The change is small, but it does suggest some arrival of ANE. Let's wait for ATP2 results, which has the best quality. If that one confirms this pattern, it might be a real thing, and quite unexpected after the findings of the last year or so. We really need Mediterranean and Atlantic aDNA from around 3500-3000 BC time frame to learn where this could have come from.
Blogger Aram said...
"BTW If You can read Russian I can send You a link where You will learn that Burzyan Bashkirs are also called Alan-Burjans because they claim to have an Alanian origin, and it is recognized by other Bashkir clans.
The amazing thing that there is another nation Osetians who claim an Alanian origin. And guess what they share the common R1b Z2103+ Z2109+ subclade. :)
The other thing with Ossetians is the fact that they don't have practicaly the Iranian Z93. The only clade that links them to the IE world is the Z2103+. The rest is the Caucasian G2a.
September 13, 2015 at 7:16 AM"
Can you post the link or a version with translated pertinent material. Ossetians have R1b-9219 the same found in Eastern Europe. There might only be 1 Armenian with that line of R1b. It is also found among Jászság people.
We also know that R1b-a2103 is about 1000 years older in the region of Sintashta. However if true about the "Alan-Burjans" then it is very interesting how it might fit in with D. Ya. Telegin observations about the Balto-Slavic ancestors; and how they may have come in contact with ancient Iranians.
"Excavations between the rivers Orel' and Samara have uncovered burials of a syncretic nature that attest contacts between the spheres of the Corded Ware and Yamna cultures. It is suggested that these may indicate early contacts between proto-Indo-Iranians and the prehistoric ancestors of the Balts and Slavs. "
@A
There was certainly some Z2103 among Scythians but it was a minor lineage and this probably also the reason why no Scythian sample so far carried it. Ossetians adopted Iranian languages later than people in Sri Lanka and are genetically indistinguishable from Caucasians so they are not representative for steppe Indo-Iranians. Z93 is the core Indo-Iranian lineage and unmixed Indo-Iranians were almost all carriers of it. Modern Indo-Iranians who don't have Z93 have in almost all cases just an non Indo-Iranian Y-DNA. But I don't think that all Indo-Iranians were Z93 and some should be R1a-Z93- and in few cases also Z2103 but let us see the facts. There is no way that an unmixed Indo-Iranian population was not dominated by R1a-Z93
It would also explain the Grugni finding of R1b-z2103 in their most ancient populations.
"Zoroastrians are the oldest religious community in Iran; in fact the first followers have been the proto-Indo-Iranians. With the Islamic invasions they were persecuted and now exist as a minority in Iran."
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3399854/
Also Lurs nomadic lifestyle, could be explained
"the majority of Lurs were nomadic herders........"
Thank youAram for that interesting study! I make a few observations!
They say that Ossetes were assimilated by Alans and received from them the present language. I cannot help but notice that I recently read that the presumed Alans from Saltovo-Mayaki culture turned out to be G2 and 1xR1b and 1xN1c.
In this study there are two R1b clusters: Digor and Lezghi cluster. Tabassarans are not included in the paper, so I presume that Lezghi R1b is the same as Tabassaran R1b and Avar R1b is not recognised as forming its own cluster. On the basis of FamilyTree Ossete project, Ossete R1b is Z2105+, CTS9219+ etc. Igmayka explained on Anthrogenica that “YFull needs one or two examples of CTS9219+ to create that branch. Right now, YFull only has its parent clade, CTS7822 (a.k.a. Z2110)”. So, to me it looks like Lezghi R1b is somewhere near the root of R1b-Z2106 and Ossete R1b is under R1b-Z2110 on the same branch with European samples concentrated in Bulgaria. It looks like Bashkir R1b is on a sister branch of R1b-Z2110, i.e. on R1b-Z2108.
Digor Ossetes have 15.7% R1b and Iron Ossetes 2.6%. Iron Ossetes have 72.6% G2a1a and 11% J2a and Digor Ossetes 56% G2a1a and 5.5% J2a. Lezghins have 30% of R1b and 44% J1*, Avars 15% R1b and 58% J1* and Kubachi and Kaitak between 0 and 6% R1b and 85-99 J1*, and Chechens and Ingush are ydna-wise completely different stock, as Ingush are 87% J2a4b and Chechens 50-60% J2a4b and 16-21% J1*. It is very interesting to compare their respective ydna clusters using the genealogical mutation rate. The oldest of all clusters is Lezghi J1* (2300±700) and Lezghi R1b 2000±700 and Ingush J2a4b (2000±1000). It is interesting that the haplotype L3-M357 is almost only found in Ingush and Chechens and the founder haplotype is also found among them. The age of Digor R1b is only 800±300 and their G2a1a is clearly older, 1300/1400±500. The main yDNA of Shapsug, Abkhaz and Circassians is G2a3b1 with an age of 1400±500. The oldest yDNA cluster of Dargins is R1a1 with an age of 1500±800. The oldest cluster of Dargins and Kubachi together is J1* with an age of 1000±600.
What is my conlusion: the oldest yDNA’s in the area are J1* and R1b. J1* came from the south and R1b from the north. I agree that maybe the Yamnaya R1b-Z2106 language is extinct and maybe the extinct Hurrian/Hattia/Urartuan language is the lost language of J1* or maybe the southern Yamnaya language was close to one or several of these languages. However, it looks like the Lezgic languages are the result of the fusion of R1b and J1*.
Ingush J2a4b is the next oldest. It probably entered the area from the south. Ingush also harbour L3 haplotypes. Their genesis looks very exciting.
It is surprising that Digor G2a1a and North West Caucasian G2a3b1 are not any older. I do not know where they came from but they must have contributed to the genesis of North West Caucasian languages.
Digor R1b is not old in Caucasus, so it looks like it came recently with Alans, and as a further confirmation this branch looks like having entered Europe from Bulgaria, i.e. land of Bolgars. Moreover, Central Asian Bashkirs sit on a sister branch of Ossete R1b-Z2110.
As for Dargins, surprisingly, it looks like they were orginally R1a1.
@A
West Iranians are one of the Iranian populations today with least NE European DNA and most Pre-IE DNA. Zoroastrians in Iran are not like Indian Brahmins direct descendants of Bronze Age Indo-Iranians and neither is modern Zoroastrianism a very archaic Indo-Iranian religion nor based on a caste system like Brahmanism. Z2103 was very early absorbed by Indo-Iranians maybe already in Abashevo but please don't exaggerate it. It was never a core or dominating lineage among any non unmixed Indo-Iranians but it was probably present in low frequencies and in some cases it became more widespread because of founder effects, the assimilation of local Z2103 or simply because of bottlenecks
Coldmountains said...
"@A
West Iranians are one of the Iranian populations today with least NE European DNA and most Pre-IE DNA. Zoroastrians in Iran are not like Indian Brahmins direct descendants of Bronze Age Indo-Iranians and neither is modern Zoroastrianism a very archaic Indo-Iranian religion nor based on a caste system like Brahmanism. Z2103 was very early absorbed by Indo-Iranians maybe already in Abashevo but please don't exaggerate it. It was never a core or dominating lineage among any non unmixed Indo-Iranians but it was probably present in low frequencies and in some cases it became more widespread because of founder effects, the assimilation of local Z2103 or simply because of bottlenecks "
Sounds good. Have you noticed R1b z2103 predates Sintashta R1a by 1000 years in the region and is also found among Albanians and or Aromanian/Vlach populations. How can this be?
Coldmountains said...
I almost forgot about R1b-z2103 among some other groups like Armenians.
Also what do you make of the R1b-z2103 among archaic Greek colonies? any thoughts
http://www.biomedcentral.com/1471-2148/11/69
@A
This is true. Sintashta-like people replaced Yamnaya-Z2103 there and originated probably in Abashevo. It is certain that Yamnaya subcultures of the Volga region influenced Sintashta and i am sure that some Z2103 was assimilated during this process but hard to say how frequent it became among early Indo-Iranian communities. It was not found yet among ancient samples associated with Indo-Iranians so it was at best at best a minor lineage among this cultures but it will be likely found if much more get tested. Z2103 in West Asia and Central Asia is a mystery and some of it could be really brought by Indo-Iranians or by Turks, which picked it from Iranians, but it could also originate from migrations of people which were Afanasevo-like
@A
Z2103 in the Balkan and the Caucasus is probably a totally different story than in Asia. I think that it was spread by Proto-Albanian, Proto-Greek and Proto-Anatolian speakers.
@Manu
Portuguese does have some influences that appear to be recent and coming from French. This probably happened after the Napoleonic and the Civil Wars (called 'Liberal Wars' here). After these events everything from France was appreciated, especially where the power was centred, in Lisbon. This is probably why Standard Portuguese - and by "standard" it's meant the Lisbon accent - has some characteristics that are absent, or uncommon, in the traditional North, especially in rural areas.
In these areas 'v', 'r' and 'rr' are spoken pretty much as they are in Spain, whereas in the standard accent 'v' is actually 'v', 'rr' is guttural and 'r' can be guttural or not (it depends on the case). The place where "guttural r" is more common is around Setúbal, some people ONLY have the guttural sound for 'r'. The result is pretty funny and worth a giggle. Or a kick in the teeth, depends on the mood I guess.
However, in my opinion it's Castillian that changed a lot, and not so much Portuguese. Ladino sounds much closer to Portuguese than it does to Castillian, eventhough it derived from the latter. Here's a link to hear it: https://www.youtube.com/watch?v=6q323m0RRPU
In my opinion what sets Portuguese appart from other Iberian languages is it's nasalization, but that exists both in the North and South, even if the same word is said differently, for instance any word ending in -ão. In the North the 'a' is almost dropped and "não" ends up sounding like 'noum' (it's the 'o' that gets nasalized, not the 'a'). It might might be because the -on/-om ending was dropped later than it was in the south. One could argue that it's not totally dropped out yet, but with the massive centralist policies of the country, and the Lisbon accent taking over, these differences are getting smaller and smaller.
There is also the abundance of the 'sh' sound in Portuguese (and Galician), so much that Brazilians say we "whistle a lot". I suspect that this might be a relic from the Suebi period, but I have no solid proof. Modern Schwäbische also has 'sh' sound before certain consonants, much like Portuguese does, although we also use it on pluralization. In Northern rural areas sometimes this even happens before vowel. For example around Viseu the name of the city is usually said 'bisheu'/'visheu' (you have both 'v' and 'b', it really depends on the individual). Apparently when I was a kid I did this a lot. Yes, my family is from the North, but I live in Lisbon.
@ Maju
I agree with you about the K-A culture. I'm sure several language were spoken. You're also correct that European farmers were less linguistically diverse than West Asian. But they'd still have been diverse. And again I highlight that, despite their overall genetic homogeneity, these people came from places as far and wide as the NE tip of Africa to Anatolia. There is no way they only spoke one (Vasconic) language. Add mixture with European foragers, natural splitting / fissioning once within Europe, and you'd have a rather diverse pre-indo European landscape.
@ Aram
You are right to take a critical perspective of ethnographic evidence. For sure, what works for ethnographic examples in America and Africa does not necessarily work for Copper Age Europe. But what matter sis the repeated patterns observable - dependent on factors like geography and economy, and how they shape language diversity. So for me at least, I'll happily take the esteemed observations of Robb and Nichols. There is no reason to think that the pontic steppe was somehow exceptional, although it appears you always revert to this special thinking.
@Mike: What I'd like or dislike is not the matter: I'm much more open-minded that you seem to imagine, just that I also have a rather well substantiated ample vision, so I'm often a step or two ahead. I can't of course expect everyone to be able to keep my pace and then of course sometimes I can also "go ahead" in the wrong direction.
What matters is how can you imagine that those extremely homogeneous peoples of the early Neolithic could become so differentiated in their language. You have not proposed a model, let alone backing it with evidence. You are just saying your line and backing it with nothing, "just because" is not an argument and I feel you are wasting my time and that of other people.
@Alberto: what all those figures suggest is that we are "splitting hairs", because the Sardinian and Basque scores are very similar all of them. Some samples (La Mina's "Spain_MN", also ATP6 in this study) leans very slightly to Sardinians, others (most ATP) lean very slightly to Basques. The obvious explanation is that all them are intermediate, transitional between EEF and the Basque-like Atlantic profile, yet to be properly studied.
I've been commenting in some of the related discussions, that the Early Neolithic mtDNA data and the Chalcolithic LCT one in the Basque Country strongly suggest a relatively sharp cline, with some complexity, in the barely 150 km of the affected geography. Notably Early Neolithic Basques show three different populations: a high U* one in the Ebro, which seems partly intrusive and may relate to these ATP and La Mina samples, a very "modern" pool in the piedmont (Paternabidea, near Pamplona) and what looks like 100% WHG pool in the coast (but few samples). The LCT data (Ebro sites) strongly suggests two distinct populations living side by side in the Chalcolithic, the largest one was LCT-negative (homozygous), much as we have seen in all EEF and most MN samples (as well as Kurgan ones) but the smallest fraction was LCT-positive (also homozygous with only one heterozygous exception). This again supports a cline or the idea of various populations with (partial but notoriously) distinct ancestries in this 150 Km narrow stripe of land, largely mountains.
Hence the overall data of the Basque-plus region seems to indicate that various somewhat distinct populations lived almost side by side, some were intermediate between Basques and Sardinians but in many aspects like EEFs (notably LCT-negative, low mtDNA H), while others were more like modern Basques (LCT-positive, high mtDNA H) and even at some point there was probably a nearly pure WHG-farmer population in the coastal regions but is now extinct (has been assimilated into the Basque pool).
Notice that the LCT trait is also common in Gokhem, which is the only properly studied Atlantic Megalithic sample so far.
As for "ANE", I'd be most cautious. Something that is striking in this study is that Ma1 does not anymore appear to influence Europeans but rather that Europeans (proto-Motala, proto-SHG) appear to influence Ma1. If this is correct, then all the ANE buzz may be a mirage. There may still be something to it but everything seems to need reconsideration. It's not the first time we criticize the use or rather the abuse of ANE but this bit in this study seems to underline that even more.
@André: The woman you link to has a very striking Portuguese (even Brazilian almost I'd dare say) accent. Hence I looked at other Ladino speakers and found more neutral accents, they still tend to Portuguese (or is it Leonese?, or even Navarro-Aragonese, cf. "Morenica", where the -ica diminutive is clearly not West Iberian but typically Navarro-Aragonese). My impression is that there is a bit of a blend of the various Ibero-Romances of the XIV century, with clear archaic features that have been more clearly lost in Castilian (which is often said to be more "innovative"). The man I watched seemed to use a simple 5-vowel system, unlike your woman, however some consonants (g/j, z/c) tended to have "exotic" pronunciations; the latter is maybe archaic, and preserved in Latin America and some Andalusian dialects, but the former is totally odd and could be an influence of other contacts, losing through the centuries of isolation the ability to pronounce some of the most strange Castilian consonants (assuming that the root is actually Castilian and not a blend of various Iberian romances, which may be closer to the truth).
Interesting debate but I'd rather not get lost in it too much. It was mostly Alberto's question.
@ Maju
" I'm often a step or two ahead people"
If you say so :)
Yet you're stuck on Gimbuta-ism from 1960s, Vasconism, and seem unable to grasp the apparently abstract concept that broad-scale biological homogeneity doesn't equate with monoligualism. You pontificate about lack of references when they're staring you in the face, whilst at the same time you're happy to pontificate on the basis of your own "private research" and questionable scholars (Venneman) . So please, come down off your pony
Anybody,
Just to be clear for the Moment,
Is ATP3 confirmed to be M-269?
Do it show any ANE or EHG type elements?
a
The best and recently updated study on Caucasians is the Balanovsky et al.
You can find the STRs if You are interested in the Supplementary file.
http://mbe.oxfordjournals.org/content/28/10/2905
As far as I know there is not a specific Ossetian study. All we know about subclades from DNA projects is that the bulk of Ossetian R1b is R1b-CTS9219 > Y5587+
https://www.familytreedna.com/public/ht35new?iframe=yresults
@Maju
Yes, we're obviously splitting hairs when comparing these samples to other Neolithic ones. The difference is pretty small, and maybe just due to the low quality of the sample. But the thing is that while it might just be a hair, it's a blond hair where we previously only found dark hairs, so it still looks like something new (by a hair). If it's real or not, I'm still cautious about it.
The Treemix in the paper with Motala contributing to MA1 just shows the limitations of Treemix, in my opinion, not much more than that. ANE is still something real and that was missing in WHGs and EEFs alike, but not in modern Basques (or even Sardinians, to an even lesser degree). We can argue about where, when and how it came to Basques, but not about its presence in them.
And what these samples might be suggesting is that it arrived in a different way than it was looking after the findings of the last year or so. That is, unrelated to the steppe. Did it arrive by sea? Was it hiding in some unknown population that was already in Western Europe since the Mesolithic or earlier? I have no idea. But that's why I'm interested in finding out more about the samples: first to know if it's real, then to try to find an explanation to it.
@Andre
Thanks for your comments about this. Yes, it was mostly my question that Maju was answering.
I agree with all the features about Portuguese that you point out as differentiating them from Castilian (though I was unaware of the regional differences, I've been more to the south than to the north, and not that much anyway). And for all I've heard, yes, i think that Ladino did sound closer to Portuguese. And this might have to do with the Basque phonetic influence in Castilian. But then again, I'm not sure if Latin itself sounded closer to Portuguese than to Spanish. I'd rather think the opposite is true.
But even beyond those features that you point out: the abundance of "sh" sound (this really stands out), the nasalisation and the guttural "rr" (the "v" sound is not really important), there is something else more difficult to explain that makes Portuguese sound different. Spaniards can read Portuguese and understand it quite well (better than Italian), but listening to it is difficult to understand, while Italian is easier. These more subtle features are probably also why Portuguese people can pronounce English noticeably better than Spaniards or Italians.
In the Balkans (where I'm living now), I've noticed some common phonetic traits with those of Spanish and Italian (and Basque). Slavic languages do have a lot of "sh" and several other related sounds that are missing in the former languages, but the rest of the sounds and the way they are pronounced here is quite similar (to a degree) to the Italian/Spanish way. In Greek it's even more obvious, I think. I'm less familiar with it, though (just listen to it while on vacations), but even simple words like "Parakalo" (Please) or "Kalimera" (Good morning), pronounced just like an Italian would, can hint those similarities.
This is all just casual observation from a non-expert, of course, and might just be coincidence. But it made me wonder if it could be due to some common substratum that left a bigger imprint in phonetics than in the languages themselves. Just thinking aloud, probably.
Off topic, but: In my impression I've always found Greek somewhat similar sounding like Castilian, because on the one hand it has the "th"-like sounds, which are written as "c" and "z" in Castilian. For example Zaragoza is pronounced like Tharagotha. And then they also share the "kh" sounds, written as "j" in Castilian. Both types of sounds are not very common in languages in general, and they are missing in most other Mediterranean languages. Italians have a very hard time pronouncing these. But I think the sharing between Greek and Castilian is coincidental. Portuguese on the other hand reminds me of Rhaeto-Romance plus the nasalisation. It's by the way interesting that a guttural kind of r is also found in Welsh - maybe it's Celtic? I don't think that South Slavic languages sound much like Italian or Spanish. They do sound different from the northern Slavic languages. Somewhat clearer, and in my subjective view more beautiful. I've had the impression of a phonetic similarity with Romanian.
It also has to be said that this talk about the sound of languages partly transcends what can be captured by written language. For example, a Latin text read with a German accent sounds completely different to what it sounds like when read with an Italian accent. And I fear the correct accent cannot be reconstructed 100% for many extinct languages.
Funny sidenote: According to racial theorist John Baker, the Nordid race has a natural ease at the pronunciation of "th"-sounds. Ridiculous, if you think of the way many Germans struggle with it. Like in hypercorrect mispronunciations of the name Elizabeth as Elithabess, etc. :D And then the ease with which it's pronounced by Greeks and Spaniards...
In fact, "th"-sounds (dental fricatives) are even part of the standard Arabic inventory.
So, if Anatolian farmers are G2a, J may have come from the Iran area. Now after reading that Balanovsky paper I have a much clearer picture. However, I must make a few clarifications to my earlier post. It is possible that Chechen J2 (which developed into J2a4b) and L3 arrived from Iran to Caucasus. Georgians are mainly J2 and G2a, which could mean an admixture between Anatolians and ancient Iranians. J1 is more typical of Arabia and the Levant and their surroundings. Maybe Sumerians will turn out to be J1.
Digor Ossete G2a1a and J2a look like having a Georgian origin which is easy to explain. In order to understand the origin of North-West Caucasian groups/languages, we should trace the origin of G2a3b1. It is interesting that G2a3b is also found in Europe. Any ideas?
I repeat that among North-East Caucasian groups, Lezghis are not the only ones to have significant R1b frequencies. The highest frequency (68%) is in Bagvalals who speak an Avaro-Andian language. Avar R1b frequency is 15%, and, apart from Dargin group, other main North East Caucasian groups, Tsezic, Lak and Khinalug frequencies are not available. Tabassaran R1b frequency is 40% and they belong to the Lezghic group. On the basis of ancient yDNA, North-East Caucasian and Armenian R1b stems from Yamnaya. If it came through Ukraine, it may have belonged to a Balkanic or Ukrainian language group whatever it was. In any case, North East Caucasian groups merge European and Near Eastern and South Asian ancestry components (http://dienekes.blogspot.be/2011/09/caucasus-revisited-yunusbayev-et-al.html). You may want to have proto-Hittite being spoken in Southern Yamnaya 2300 BC but I still do not believe that IE languages were spoken everywhere at that time. J1 is even a more important haplogroup in North-East Caucasians, so when we eventually get ancient J1, it will also clarify the origin of this language group.
Moreover, when we know Sredny Stog, Catacomb yDNA and Majkop yDNA, we can make even much better presumptions. I may even change my mind! :-)
@Simon_W
Yes, we might be going a bit too off topic here, but I agree with your comments. Good to see that you do notice these traits (that do go beyond what can be written, even with phonetic transcription, it's a more subtle thing).
From the South Slavic languages, I think Macedonian is the one that sounds more similar to Spanish/Italian. Of course, to a degree, and with some obvious differences. But it's that "somewhat clearer" sounding that you mention (compared to its northern cousins) what is intriguing. And for example, Spanish is difficult to pronounce correctly for many foreigners. A Spaniards can easily imitate the accent of a Frenchman speaking Spanish, of an Englishman, an American, a German, etc... But if you ask a Spaniard to imitate the accent of a Serbian speaking Spanish he will be puzzled. Serbians (and even less Macedonians, though these are less known) don't have any special accent that you can imitate. They speak a Spanish that sounds quite close to native.
I didn't mention Romanian because I'm less familiar with it, but it does sound similar to South Slavic languages, though even a bit "harder", more Latin-like (though I'd need to hear it more to really know).
If all this has any meaning at all I don't know. Might be coincidence (the "th" sound probably is, and it's even missing in American Spanish). But there it is.
Kristiina
Baglavals number in the 2010 census counted only 1500 people. That is very small to be a source of reliable information. Tabasarans number is 130.000. Avars number is more than 1 million.
Yes J1 and J2 coming from Iran is very plausible. Btw there are theories and claims that Avars came from Khorasan. I am sceptic of that but it is not improbable. https://en.wikipedia.org/wiki/Avars_%28Caucasus
I also think J1 is the best candidate to be a Sumerian haplogroup.
Returning to R1b-Z2103. It is important that the basal and unique subclades of that branch are present in Dagestan. This is a serious indication that Z2103 moved to South from Steppe trough the Eastern side of Caucasus. Also it is possible that in Bronze Age the Caspian Seas was much lower and that corridor was much wider.
The archaeological data also confirms the apparition of new type in South Caucasus somewhere around 2700-2500 BC.
@Mike: ... "you're stuck on Gimbuta-ism from 1960s, [Venneman's] Vasconism"...
Neither one nor the other. That's just something you say to try to disqualify my opinions, dirty debate tricks that go nowhere and may actually harm your own credibility if anything, at least in my eyes.
"... and seem unable to grasp the apparently abstract concept that broad-scale biological homogeneity doesn't equate with monoli[n]gualism".
This is not "broad-scale" anything: the EEF people were genetically almost "identical", like one Sardinian from another. There's a bit of change within the western samples, notably as time passes ("MN") but we can identify it all with WHG (or other HG) admixture and it never is as large as to suggest that the HG languages became dominant.
I can understand that it may be debatable but you are not debating, just denying upfront.
@Alberto: I have argued in the past that ANE is a mirage, mostly because it is too old to be really meaningful, and asked for more specific references such as EHG or Yamna. Oddly enough I have also been skeptical of Treemix in the past but in this case it seems to add weight to my pre-existent ANE-skepticism, so...
The Günther & Valdiosera trees (which are both consistent for this issue) suggest that what Lazaridis called ANE may be just SHG/EHG "Eastern Gravettian" stuff that is present in Ma1 rather than an autonomous real component. I think we are much better off forgetting about ANE, which in Lazaridis it was clearly just identifying (rather imprecisely) Eastern European and Caucasian elements that were not present in the other two ancient references (WHG, EEF). ANE is surely a proxy but, as all proxies, it can be misleading, especially because it is not too close to the real thing.
It's not that ANE is "not real", just that it is not real enough to be of any quantification use. We need to stick to Holocene references, of which we have already a bunch (no need for ANE hence).
Using these Holocene references, we get two (or more) very different results. Haak & Lazaridis suggest a major Yamna impact, even in Basques, while Alentoft instead suggests a much weaker and clinal impact, reaching zero in Basques. This last paper, while not using Eastern European references, does not seem to support major steppe impact in Basques, just residual.
So what do we do when independent studies produce different results? Raise an eyebrow and wait for improved ones and/or resort to complementary data such as the one we can get from archaeology. And this last is what brings me dramatically opposed to Haak & Lazaridis' extremism, because the only possible vector of the alleged massive demic change in Western Europe would be Celts (Urnfields, Hallstatt and La Tène) and these cultures in most cases do not show indications of such aggressive mass replacement, or so I understand. Hence I suspect that a much weaker colonization drive, maybe complemented with pre-IE remixes that we don't yet understand (largely for lack of Atlantic aDNA), is what most likely happened.
My opinion in any case.
@Simon: the "kh" sound is probably an Arabic borrowing. As for "th" it may indeed be Atlantic or Nordic, I can't say (continental German may not be a good representative). It'd be interesting to consider here the Celtiberian substrate also, unsure how this works in Celtic in general. What is clear is that neither sound is Basque.
@Simon again (PS): After checking, both sounds are common in Semitic (either Arabic or Hebrew) but are also found in Celtic notbaly. However while "kh" is common in Gaelic, "th" is only found in Brythonic. So either Celtic substrate or Semitic influence (or both reinforcing each other).
Interestingly, voiced and voiceless "th" sounds are also found in eastern Emilian and lowland Romagnol - might really be Celtic. These dialects are also very rich in vowels (about 20, while standard Italian has only 7), like some modern Celtic languages also.
@Maju
Pure ANE never spread much, so isolating it in admixture is an interesting exercise and sometimes quite useful. But in other cases it's better to refer to real populations, like Yamnaya or Caucasus ones, I agree.
But the results won't change much when either method is used correctly. The two above examples that you mention (Haak et al. and Allentoft et al., and I think you refer to the admixture run in Allentoft and to the Yamnaya estimates graph in Haak et al.) are not too correct for similar but different reasons. One needs to understand how the methods work and contrast them with many other admixture runs, formal stats, IBS, etc... to get a good understanding. I've dedicated time and effort to that specific task, for a few years. Still learning, though.
The idea that Basques are a pure Neolithic population without any "eastern" (be it Yamnaya or Caucasus or EHG) influence is not correct. But the idea that Basques have less influence than all other modern Europeans except Sardinians is correct.
So I really don't know why or what we are discussing here. If you are arguing that Basques have 0% of these eastern genes, it's wrong and it can be proved with any simple formal stat, for example of the type:
D(Mbuti, Yamnaya)(Spain_MN, Basque)
It will show that Basques are closer to Yamnaya than Spain_MN. You can change Yamnaya for EHG or for Georgian, the result will be similar.
Now, when and how this eastern influence arrived to the Basque Country (and Spain, and maybe all Western Europe) is still unknown. Up until now (based on the findings from the last year or so), it seemed that Yamnaya was the most likely source. But now these samples from Atapuerca could indicate a different time and source (still unknown). And again, I agree that the change between these Atapuerca samples and the previous Spain_MN ones is probably very small. But if it's real (no matter how small), it's already a very interesting finding. Because the change did happen at some point. So getting a clue as to when and how it happened is a step forward.
If the results from these samples turn out to be spurious, well, then we'll be at the same point as before. But for now I'm interested in seeing if they are real or not, because I think it can be important.
FYI, regarding Romagnol, there you find some audio files of Romagnol poetry:
http://www.homolaicus.com/arte/cesena/storia/poeti/Galli/sommario.htm
To me it sounds not too distant from Rhaeto-Romance, or more precisely Romansh. Which makes sense in the light of modern attempts to systematise Romance languages. Rhaeto-Romance probably isn't a valid category, but there is a North Italian group of dialects that includes Romansh and Gallo-Italian variants like Romagnol, among others:
https://upload.wikimedia.org/wikipedia/commons/thumb/3/31/Romance-lg-classification-en.png/1024px-Romance-lg-classification-en.png
@Alberto: "It will show that Basques are closer to Yamnaya than Spain_MN".
Necessarily so. There are two reasons why:
1. Basques are not a fossil nor have existed in a safety box perfectly isolated from other populations for millennia, so some IE admixture has necessarily infiltrated the Basque genetic pool. However in most many analyses it seems pretty much negligible, not in Haak's however what rather seems to bring all the Haak's analyses into question, because independent analyses don't seem able to confirm his "findings", what is a typical indicator of hidden bias.
2. Basques are significantly more HG than Spain_MN, and that makes them closer to Yamna or EHG.
The question is to first differentiate between various possible HG increment sources, such as WHG, SHG, raw EHG and Yamna (can't think of any other). So you would have to repeat the same formula with all four possible extra HG sources in order to discern which is the best match for that extra HG in Basques. In order to be safe I'd even differentiate in WHG between Braña and Lochsbour and maybe even between Motala and Ajvide as well.
Ideally you could even run ADMIXTURE in supervised mode with several pre-selected ancient samples as forced populations but, for what I've been told, their coverage is often not good enough for that.
PS- Regarding point 2, second paragraph, of my previous comment...
When we look at the Europe-only PCAs, the PC2 is invariably polarized Malta/Caucasus vs Basques. This clearly indicates that when you say: "You can change Yamnaya for EHG or for Georgian, the result will be similar", I can say with huge levels of certainty that you are wrong at least for Georgians (or any other Caucasian, like Lezgins).
In those PCAs it is apparent that what all that "Mediterranean" axis of the "European cross" share is a deviation from the Sardinian (or generic Neolithic cluster) baseline towards NE Europeans (or generic HG tendency) but there is no tendency towards the Caucasus or Eastern Mediterranean, certainly not in Basques who clearly show the opposite polarity, more marked than any other Europeans. This is less clear usually in all West Eurasia PCAs because PC1 is hijacked by the Asia vs Europe polarity and PC2 by lowland vs highland West Asian internal distinctions, at least partly.
So Basques have more HG than Neolithic/Sardinian samples but less Caucasian than any other European population.
@Maju
I still don't know what is it you are discussing here. Is it about the quantity of "Indo-European" blood in Basques? I'm not discussing that. With what we know today, it's about 20%, but that's not what I'm debating. If you want to argue that it's 10% fine, it's not what I'm interested in right now.
What I'm interested in is finding out when, and from whom, this genes arrived to Iberia (let's just forget about Basques and refer to Iberians, that way it will be easier, it seems).
Up until now, Yamnaya migrations was the best hypothesis. But these Atapuerca samples suddenly hint that maybe it was not Yamnaya. This is the interesting part. It's not about Basques. It's about Western Europe in general, and the expansion of ANE and R1b in Western Europe (whether connected or not, that's another interesting thing to find out).
BTW, I like David's latest K12 here. For European populations it works well and it's quite informative, without any real flaw (probably for others too, but I didn't check those too carefully). You'll see that Basques don't have significantly more HG ancestry than Spain_MN, but roughly the same. If I have the chance I'll try to get that D-stat comparing Basques and Spain_MN to Georgians.
That is important because when (Haak & Lazaridis) the estimated IE amount of ancestry in Basques gets higher, it gets also higher (and flatter) across the board, while, when it gets lower or zero in Basques (most other analyses), it gets lower (and more clinal) across the board. And IMO the second solution is the only one that makes sense.
So I'm interested on the Basque reference not because of Vasco-centrism but because it is very symptomatic and informative of what happens around as well. Also you mentioned that you can exchange EHG for "Georgians" and I say that nope, you can't: repeatedly the data shows that Basques are less Caucasian than the Caucasian affinity is transversal and not directly related to HG/Eastern European affinity, peaking towards the Eastern Mediterranean.
It's not these Atapuerca sample is the extremist reading of the Yamna model, such as the conflating of "the expansion of ANE and R1b in Western Europe", what is clearly two different things (assuming ANE is a "thing" at all).
"Basques don't have significantly more HG ancestry than Spain_MN, but roughly the same".
Very very roughly only: there is clearly some notable more HG, ~10 percentual points what implies extra 20% admixture with hypothetically pure HGs above a La Mina or Atapuerca type baseline*. This extra 20% (and whatever is associated with it) drives Basques towards Russians relative to Sardinians and the Neolithic cluster. While Spaniards align with Basques in this displacement along the European PC1, their displacement cannot be attributed to extra HG instead but rather only to whatever Caucasus+ came from there.
So these are almost certainly two different vectors but we are failing to discern them, partly for lack of ancient Atlantic samples.
___________
* We don't only have to account for the HG that overlaps the baseline's Neo fraction but also for what overlaps their HG fraction. In the baseline samples that is roughly a 1:1 apportion (per this last study), so double.
Having this in mind is most important because you also have to consider what HG fraction comes from the East with the Caucasus-Baloch component, etc. Not all the "blue" segment is the same and there probably lays the root of the misunderstandings.
Erratum: ... "Basques are less Caucasian than the Caucasian affinity is transversal"... should read ... "Basques are less Caucasian than any other Europeans and that the Caucasian affinity is transversal"...
@Maju
Chad and David were kind enough to run the stats:
Spain_MN Basque_Spanish Yamnaya Mbuti -0.0106 -2.686 64540
Spain_MN Basque_Spanish EHG Mbuti -0.0096 -1.807 63527
Spain_MN Basque_Spanish Georgian Mbuti -0.0064 -1.845 64627
Spain_MN Basque_Spanish WHG Mbuti 0.0008 0.164 61523
As you see, all of the 3 above (Yamnaya, EHG and Georgian) are closer to Basque than to Spain_MN. This is something extremely obvious and I have to wonder why on earth you could think otherwise. Please look at the values of the K12 I linked above for these 3 populations: Basques, Spain_MN and Georgians, and then ask yourself if you have one single reason to think that Georgians could be closer to Spain_MN than to Basques.
I really hope that I'm not sounding patronizing, but you really have to understand that basic reality because otherwise you will continue to say "with a HUGE degree of confidence" things that are simply wrong (and very basic, really).
The last of the stats show, as expected, that WHG has higher affinity to Spain_MN than to Basque. This does not mean necessarily that Spain_MN has more WHG admixture than Basques (they have a hair more, but just a hair). It's also because Basques have some 7% ANE that Spain_MN lacks, and that small amount is enough to turn the stat on the side of Spain_MN, if not by much.
As for the comments you make about admixture runs and the Caucasus component in Basques, I wish we could sit together one day, with pencil and paper, a few examples and simple graphs, because that way i could explain you very easily why it is and you will stop being confused by these seemingly contradictory results that you mention. But here with words it would be difficult and long, and off topic. For now I can just advise you to bookmark the K12 above for reference, because that one is quite good in this respect and won't lead you to confusion.
The only thing you got right is that Basques have less Caucasian/Eastern/Indo-European blood than other European population (except Sardinians). But that's something we all know well and we all agree. What you have to finally understand is that they have more than EEFs (or MN ones), including the Caucasus component.
Finally I will say that I have much respect for your knowledge and your opinions in most of these fields, and for you personally. But with a few more technical things that you don't have a really good understanding of, you can't argue with that huge degree of confidence. At least give enough room for doubt and for others to be right. Otherwise the debates become absurd and unnecessarily long and boring.
It's very possible that I'm reading too much on a simple PCA dimension, what I see is that Sardinians (and the Neolithic cluster that tends to that same polarity) are clearly more like Lezgin (or Georgian, I presume) in one of the PC axes (PC2, Europeans only), just as most other Europeans are (from Spaniards to Russians, all align at about the same Basque-Caucasus intermediate distance in that axis).
However Caucasus is still be more Basque-like (or vice versa) in the PC1 and PC1 always weights more, so, you are right. It would be different if we'd use modern Spaniards however (because Basques and Spaniards alight at roughly the same position in PC1).
This is a mystery that I have yet to properly understand but that I suspect very important in all these considerations about European origins. It can't be just meaningless that the European PCA systematically produces a Basque-Caucasus or Basque-Malta axis in PC2, with Paleoeuropeans scoring not just on top of the PC1 (Russian polarity) but also at bottom of the PC2 (Basque polarity). This last is more notorious in WHG, with SHG being more neutral instead but still somewhat slanted towards the Basque polarity and ANE being strictly "neutral" (haven't seen EHG or Yamna in Europe-only PCAs yet).
See for example this PCA from Lazaridis (barely annotated by me): the differential angle between the ANE neutrality (no EHGs nor Yamna there) and WHG is suggestive of some information we are missing here. Or also this one from the Portalón paper. Or this one from Dasakali (a bit distorted because of Levant samples' presence but includes Adigey and also Ajvide, who are becoming a hot suspect and show the same Basque-like tendency of WHGs).
In the end what we get is that there are two HG polarities (which may well include some ANE-like stuff or not): (1) EHG/Yamna vs EEF and (2) WHG(?) vs Caucasus/Eastern Med.
The fact that we systematically find some Caucasus component associated to Yamna/EHG influence is very telling but of course Sardinians/EEFs do not have it and their relative greater similitude in PC2 is because of other factors such as shared affinity with Levant.
Confounding factors criss-crossing here: it's not a mere "triangle" but something a bit more complicated and once we begin discerning more areas now unsampled for nuclear aDNA we will have a better understanding of this complexity.
@Maju
Yes, PCAs can be made using different strategies and show different patterns, but as a rule of thumb, if they show ancient genomes (pre-Bronze Age) among modern European populations, it means you can't take them literally. The one from Lazaridis showing MA1 among Icelanders, or Skoglund_farmer (Gökhem) between Bergamo and French is a good example of it. But the others too, to a lesser degree.
And I know you place more trust in scientific papers than in bloggers, and that's all right. But this is not an ideal world, and scientific papers can often be quite sloppy and contain some real "pearls". But Davidski knows pretty well what he's doing and you can usually trust his tests more than you can trust many published ones. And then the blogs have comments in real time and many well informed people can comment about the tests, etc... (don't miss what Matt usually has to say about any test: he's not only well informed and smart, he's also very objective. But there are many others too).
And now it's time to move on. I'm still awaiting for more tests from those Atapuerca samples. They got me intrigued, even if I'm still a bit sceptic about them being any different from previous Neolithic ones (it might be just noise due to low quality or some other artefact). Once we get more tests, we can debate some more about them. For now there's really not much more to say.
Alberto: "Basques don't have significantly more HG ancestry than Spain_MN, but roughly the same".
Maju: ”While Spaniards align with Basques in this displacement along the European PC1, their displacement cannot be attributed to extra HG instead but rather only to whatever Caucasus+ came from there.”
Alberto: ”As you see, all of the 3 above (Yamnaya, EHG and Georgian) are closer to Basque than to Spain_MN. This is something extremely obvious and I have to wonder why on earth you could think otherwise.”
Basques are modern people and they must have had some later admixture with Celts, Spanish etc, so they should naturally be closer to Yamnaya, EHG and Georgian compared to Middle Neolithic Spanish or ancient Basques or am I missing something important here. If Basques/their ancestors have mixed with IE’s and have received geneflow from the Mediterranean Basin in the course of the past 4000 years, it does not mean that their R1b must be of Indo-European origin.
Kristiina: Basques are modern people, so are Sardianians. However in many analyses neither shows any significant IE-type admixture. This study for example shows only tiny "black" component in Basques and Gascons, almost as low as Sardinians'. It is not the only one: Alentoft's data goes in the same line.
I think that is what you are missing.
Hence the "ANE" apparent extra detected in Basques by Lazaridis is confusing and either has some other pre-IE origins or is a mere artifact (or a mix of both). All this is made more confusing by Haak & Lazaridis' 2015, which seem to claim a major Yamna-like impact in Basques, something that is not detected in other studies like this one or Alentoft. So there is a major contradiction in the data and either someone is committing a serious error or something very hard to explain is at play.
Note: Why is that there is so little Yamna-like admixture in Basques (per most but not all studies)? Probably because the IE migrations West of the Rhine are very recent, dating only to c. 1300 (oldest, Urnfields flow along the Rhone and Ebro) and c. 700-400 BCE (Hallstatt & La Tène: Celts proper and more general migration/conquest). These triggered extensive conflicts and rather low admixture among populations that remained hostile to the Celtic advance as were Vasco-Aquitanians, and that later were not significantly affected by the Roman conquest, being too peripheral to matter (unlike Iberians for example). This last can also be said of Sardinians, even if they did lose the language, who were also unaffected by the Italic tide that parallels Celtic expansion in Italy.
Notice that the amount of Yamna-like admixture in Basques is relevant not just for Basques but also because they are pretty much baseline for other Europeans, particularly Western/Northern ones. So if Basques are 20% Yamna-admixed, then French are like 40%. But if Basques are nearly 0%, then French are only 20% Yamna-like. Rough figures anyhow.
So which is the correct data dramatically affects the conclusions for Europe in general, not just for Basques. And holistically thinking, I rather think that the reality is much closer to the 0%/20% figure than to the 20%/40% one that the Haak & Lazaridis team seems to favor. This still leaves the issue open of what happens with all that extra ANE, in Basques as in Britons?, how real it is?, if real, where does it come from?...
Maju, I agree with you. I am not so sure that all that is defined Yamnaya comes from R1b-rich Yamnaya culture located in Bronze Age Southern Russia. According to Haak, Estonians are 55% Yamnaya (Figure 3) and according to Davidski Udmurts are 65% Yamnaya. Instead, it is much easier to claim that 55% of Estonian ancestry comes from the Ukrainian Ice Age refuge but IMO Yamnaya is quite narrow in space and time.
That's not my point, probably because I'm not really bothering discerning Eastern Europe the way you do (and that's why having different complementary focuses is something good). I'm only looking at Western Europe, or Central Europe also, but not really at Eastern European nuances because what is clear for the rest is that the new elements are Eastern-like, Yamna or whatever else, something that doesn't seem to be in the Western and Central parts of Europe (barring maybe Scandinavia and unsampled areas) before that period of Corded Ware, etc.
I'm also strictly discussing autosomal DNA (maybe with a complement of mtDNA but totally ignoring the Y-DNA aspect until more data arrives). So Yamna and Corded Ware are nearly the same, even if their Y-DNA seems different (who cares?), and they are the principal vector of Eastern European type admixture in the rest of the continent. The issue is how much of that admixture is real and how much is a mirage - or maybe also real but has some other older source like, say, Doggerland-type substrate.
@Maju
The contradictory results that you mention are an artefact of the samples used in different admixture runs. I will try to explain it briefly:
Maybe of the admixture analysis use either modern samples only, or even if they use ancient ones they mix it in not very clever ways. Let's focus on the Euro_HG cluster. We know that WHGs (Loschbour, La Braña, KO1) didn't have any ANE at all. But we also know that SHGs all had some 15-20% ANE. And EHG close to 40% of it.
Davidski replicated quite well the admixture run from Haak et al., where Basques have no "teal" component.
In that run, the Blue component was a mix of European HGs, so it was not pure WHG. It actually was closer to SHG. He run the components through his on K8 and plotted them on his "Fateful triangle", here.
Now look at that plot and draw a line from the EEF (Orange) to the European (Blue) component. What happens? Basques and Sardinians stay west of that line. They don't need any "teal" to place them close to their real position. But his is because the blue component has some 18% ANE, and both Basque and Sardinians have a WHG/ANE ratio that is high enough to fit into that western part of the line. In instead the "Blue" was pure WHG, the line will leave Basques and Sardinians to the east, and they would require some "teal" to get their low ANE amounts. Compare to Lithuanians (for example), their position in the plot and their values in the spreadsheet. They need Orang, Blue and Teal to be placed correctly, because their WHG/ANE ratio is too low to fit into the west side of the Orange-Blue axis. Same for all other Europeans.
Calculators based on modern samples work similarly. They will have a "European" cluster that will peak in Lithuanians, usually. They can show maybe 90% of it (for example, Dodecad's K7b). If we take this K7b as an example, Lithuanians are close to 90% Atlantic_Baltic and 10% West_Asian (basically "teal"). But then you look at Basques (French_Basque, Pais_Vasco) and they are some 73% Atlantic_Baltic, 27% Southern and 0% West_Asian.
Why is this? It is simple. Lithuanians are nowhere near 90% WHG. The Atlantic_Baltic cluster has WHG + ENF + ANE. While the "West_Asian" cluster has ENF + ANE. And the "Southern" cluster only has ENF. And again, Sardinians and Basques are the only ones who's WHG/ANE ratio is high enough to not need any "West_Asian" (teal) to plot in their correct position. Spaniards already need small amounts of it, and other Europeans more.
It is quite a simple limitation of the way admixture works, and if you get to understand it you will also start to understand why and how each calculator behaves and why. This is a small example, but I hope enough to get the basics of it.
There is no perfect calculator as of today (we miss many ancient samples, plus most people don't even understand how to use admixture sanely), but looking at different ones, and understanding a bit how they work, which samples they are using, what does each cluster really contain (or "hide") can give you a good idea overall of what is what.
^Ŝorry for several typos above. not enough proof reading, and no edit button. Don't want to replicate now all with links, etc... but I hope you don't get confused by the typos.
With what we know today (of course, this is subject to change when new samples come that show something different), the populations that introduced ANE into Europe (or most of Europe, since there was already some in the Baltic), had about 1/3rd of ANE (either Yamnaya or the "Armenian-like" population AKA "Teal"). So the impact of these ANE-rich populations can be roughly calculated by multiplying the ANE on modern populations by 3.
Basques have some 7% ANE (maybe 8%, even). So 20% for Basques should be in the ballpark (again, unless some new finding changes our current knowledge, which is possible). Sardinians some 10-12%. Spaniards are around 30%, while most other Europeans are in the 40-50% range.
I really find it difficult that these figures will change much. It's quite unlikely that we find ANE in the places of Europe where it hasn't been found already. And it's unlikely that we find a population with higher ANE than 1/3rd that came to Europe. For Basques to have a lower than 20% impact, we'd need for example a population that was 50% ANE responsible for bringing it to the Basque Country. That would reduce the impact to 14%. But it's unlikely such population will pop-up out of nowhere.
You seem to have it more clear than I do, Alberto. Thanks for the explanations in any case but I don't get it.
Anyhow, according to Lazaridis, Basques have almost 10% ANE (vs 13% Spaniards, 18% Estonians, and 4-5% Sardinians). Even if we use Sardinians as baseline, there would be still 5% ANE in Basques vs a max. 13% in Estonians. Either way it'd mean some 40% Estonian-like admixture in Basques and more than 50% in Spaniards, what is just not credible. So ANE must be an artifact of some sort.
Later data does seem to suggest that indeed there are not such extreme levels of Eastern admixture (nor anything even remotely close) but people still talk of ANE and Yamna... And I understand that they do because I myself still haven't got any good explanation for the Lazaridis+Haak massive apparent fractions of Eastern admixture.
According to Lazaridis Table S14.9, Basques are 11.4% ANE, Spaniards 12.3% and Estonians 18.3%.
But what is ANE? It's basically MA-1, an Upper Paleolithic Siberian. Very far from Copper Age Europe, both in time and place. So it's clear that ANE-admixture shouldn't be taken too literally as ANE admixing into Neolithic Europe.
But now we know EHG, SHG, Yamnaya and BA Armenia, and they all had some ANE-relationship, as suggested by formal stats. People like these must be responsible for the ANE-affinity in modern Europe.
If pure ANE had admixed into Europe, we'd only need little admixture. If the ANE-affinity had come from an Estonian-like population, we'd need a lot of admixture. Now we know Yamnaya had much stronger ANE-affinity than Estonians, so the necessary amount of admixture is reduced again.
There was some (though not enough) ANE in SHG, and we don't know British or Doggerland HG. It's possible that they had a little bit, though I guess they hadn't. Loschbour from Luxembourg was quite northwestern geographically.
@Simon: Yamna could have some 25% "ANE", just somewhat above Estonians. So the necessary admixture would still need to be massive (>40% in Basques, who otherwise show no signs of Yamna admixture)... unless ANE is a mirage or has other sources.
Haak et al., who published the first Yamnaya genomes, abstained from calculating any ANE admixture proportions, but according to David's K8 model, Yamnaya is around 35% ANE. So this would mean about 25% Yamnaya in Basques, which is very close to the estimate in Haak et al., who arrived at 26.2% Yamnaya in figure S9.27 C.
I don't think ANE is simply noise, the affinities have been demonstrated with formal statistics.
@Simon: that's impossible even with David's data apparently (h/t to Alberto) but also with Alentoft's and with Günther's data as well. That "ANE" that Lazaridis attributes to Basques comes from nowhere similar to Yamna. Either it's a mirage, an artifact, "noise"... or has some other yet unidentified origin.
I wouldn't call it impossible. Your personal expectation is that biological kurgan influence in IE Europeans is minor admixture and much rather a cultural elite dominance effect, and hence non-IE Basques would be even less kurgan admixed. But that's just your opinion, based on archeology and maybe other factors. Let's see how that will change in the light of new archeogenetic evidence in the next few years. I'm not sure why you think David's K9 run also speaks against it. I guess you mean because of the strong teal component in Yamnaya vs. its near complete absence in Basques. But as Alberto explained, admixture runs like this can't be taken as absolute; Alberto explained that the Basques don't need any teal to explain their ANE because the blue hunter-gatherer component already has some ANE. That's because it's an average European hunter-gatherer component that also includes some EHG and SHG influence, it's not a WHG component. And for instance in Dodecad K12b Basques have quite a lot of the Gedrosia component.
@Simon: "I'm not sure why you think David's K9 run also speaks against it. I guess you mean because of the strong teal component in Yamnaya vs. its near complete absence in Basques".
Exactly.
"But as Alberto explained, admixture runs like this can't be taken as absolute; Alberto explained that the Basques don't need any teal to explain their ANE because the blue hunter-gatherer component already has some ANE".
ANE is at the moment nothing more than a disturbing mystery to me: mirage, artifact or real? Can't say.
In any case: no teal, no Yamna. That is straightforward. Some Basques do have minor teal but it's residual, what makes total sense, so I'm satisfied with this explanation.
Appealing to ANE won't solve the lack of teal, which should have been carried (more or less diluted) by any Indoeuropean genetic flows, be them Yamna or Corded. This is not a problem of ADMIXTURE, it is a problem of ANE, which behaves as a ghost and may well be a ghost after all.
Alternatively there is another source of ANE, which is pre-Indoeuropean but has not yet been detected.
But no teal = no IE blood.
@Maju,
"But no teal = no IE blood."
There are plenty of methods besides ADMIXTURE which show an eastern-shift in Basque. Haak actually modeled Basque as about 20% Yamnaya. Basque were literally surrounded by IEs right before the Roman empire.
Haak's results seem rather contradictory with other results, and also with common sense. How do you explain that, Krefter? Claiming that Haak & Lazaridis are the wise ones and all the rest are wrong, or considering that when some results are not replicable, they may well be wrong.
I just see more and more contradictions between Haak & Lazaridis and comparable independent studies, all comparable studies don't seem able to reproduce Haak's Yamna fractions in any way. Not even David, who is clearly favorable to them. Why? But more importantly should we keep arguing based on results that are not being replicated? I know of theories and careers thrashed for much less.
In any case I know well that Haak's data models Basque as c. 20% Yamna, and hence other Europeans as much higher figures. And I know that such position exists and that you and a few others are "paladins" of it.
But I also know that there many other analyses (not just one but all the others) that model Basques as near 0% Yamna, reduce the overall influence of IE (Yamna) ancestry across the board and produce a softer and more clinal such influence on the map, that makes better sense with other non-genetic data.
Which should I favor? The extremist unreplicated results, just because you are adamant, almost fanatic, of it... or the one that seems to make better sense? I'm trying to make sense of paleohistory after all.
But let's agree on this: the problem needs more research, and probably the answer will come via more samples from unexplored areas.
@Maju,
"In any case I know well that Haak's data models Basque as c. 20% Yamna"
20% Yamnaya in Basque does sound crazy, in other areas though David gets the same results as Haak and Laz.
@ Maju
David didn't include the early Bronze Age sample BR1 from Gamba et al. into his Teal K9 analysis. But that whole talk about teal people and the teal component roots in the ADMIXTURE run of the Haak et al. paper, and there you can see that one of the HungaryGamba_BA samples, namely BR1, also almost lacks teal. She has marginal teal, yet quite considerable eastern shift and ANE in all other analyses including PCA. The explanation for this is still unresolved, but to me it's most plausible that she had considerable admixture from the western part of the steppe, where the local hunter-gatherers had quite some ANE, and where the West Asian influence from the Transcaucasus was low. These people were apparently the first steppe people who entered the Carpathian Basin in any significant numbers.
But were they IE speaking despite the very low teal? Obviously it cannot be said with any certainty, but IMHO they were IE.
As you said yourself repeatedly, Yamnaya wasn't the first and only kurgan steppe culture. And while the corpus of ancient DNA data suggests an eastern shift in Europe, it doesn't necessarily suggest that all of this is from Yamnaya. It also has to be said that so far we've only got autosomal Yamnaya DNA from east of the Don - perhaps it was different west of it.
In any case, this high EHG / low teal population from the Carpathian Basin might be the explanation for David's D stats showing that ATP9 has the strongest admixture signal with EHG, not with Yamnaya or Corded people. Admittedly, Bronze Age Hungarians don't produce an admixture signal with ATP9 either, but these were perhaps already too admixed with the locals in Hungary, whereas the population that admixed into ATP9 moved on before the Bronze Age.
I think the rather late arrival of this eastern admixture in El Portalon suggests that this came in all likelhood from IEs, and that the Basque language comes from people similar to the earlier ATPs who didn't have the eastern shift, for example ATP2. But as David's D-stats PCA shows, modern Basques have an eastern shift towards other Europeans, relative to ATP2, who sits in a more extreme position:
http://2.bp.blogspot.com/-tLc8zG4kvQk/VfkkEnimY6I/AAAAAAAADZw/6zc_ndYgBVM/s1206/West_Eurasian_PCA_ATP2.png
So IMHO this shows that Basques acquired some admixture and yet retained their language. This reminds me of the Armenians, who had a similar population history. The Iron Age Armenians were quite distinct from modern Armenians, because of additional Near Eastern admixture in the latter, yet they retained their language and ethnicity which was constant at least since the Iron Age.
I think ANE isn't that mysterious, it's an Upper Paleolithic genome from Siberia (or two, if we take AG-2 into account) which shows affinity with a number of modern and ancient populations. ANE affinity hence may in principle come from various sources, but in Europe in general it means eastern admixture.
Günther et al. presented formal stats according to which there was affinity with SHG in some EEF. This may well be the case, but the fact remains that these EEF lack ANE affinity, unlike the SHG. So it's probably not direct SHG admixture, but rather some relationship between some WHG and the SHG that goes beyond the ANE in the latter. (And probably farmer admixture in the later SHG also contributed to the affinity.) And hence it can't explain the ANE shift in LNBA and modern Europeans. I would also point out that they didn't include EHG in their analyses.
@Simon: Indeed the Hungary Bronze Age sample (Vucedol culture presumably) is very odd, much like the Baalberge ones, although I did not know it produces a high ANE signal. Your interpretation however should be tested in a WHG-EHG-EEF triangle model, only if EHG is high enough to correspond to the ANE signal, what you say would make sense.
But yours it is a good idea to be tested, and also for all the "MN" samples, and even modern Basques.
One problem we may have is the issue of Nordic HGs, which I suspect at the origin of these issues since long ago. Günther & Valdiosera get a surprising 40% pre-SHG (Motala) in Mal'ta AND also strong admixture from pre-Ajvide into Gökhem. Ajvide and the other Pitted Ware samples should have roots in Eastern European EHG-like populations, although maybe mixed with SHG type peoples. A major question is was there a strong "ANE reservoir" in NW European hunter-gatherers, additionally: did Pitted Ware brought pre-IE Eastern genetics (incl. ANE but not the Caucasus or "teal" component) as far as to the Atlantic? Of course a related but maybe not the same issue is if Vucedol people were more or less like these.
... "modern Basques have an eastern shift towards other Europeans, relative to ATP2"...
But ATP2 and other ATP samples did not even show lactose tolerance and also had a "pre-modern" (low H) mtDNA pool. Meanwhile at least some Neo- and Chalcolithic Basques had these "modern" traits already, as does Gökheim and (mtDNA only) Gurgy. So the Neo-/Chalcolithic Atlantic holds some fundamental clues that ATP cannot provide in spite of their transitional nature. The shift probably existed to some extent already in ATP2 times.
Actually I classify pre-IE farmers as follows:
1. First farmers (EEF, Sardinians), with around 25% Paleoeuropean maybe, mostly from the Balcans.
2. Transition zone farmers (ATP, La Mina, etc.), with some extra 30-35% Paleoeuropean incorporated at destiny or on the march. But still lactose intolerant and with a pre-modern mtDNA pool (although they often have high frequencies of U not really of H)
3. Atlantic farmers (modern Basques would be close to the archetype, Gökhem should also be considered in this cluster even if they show some type 2 traits, at least some ancient "Basques" like Paternabidea or the LT+ Chalcolithic ones, Gurgy probably too, etc.): with significant or even fixated LT, modern mtDNA pools (high H but lower U) and a big question mark re. everything autosomal but between the ~35% of extra HG of Gökhem and the ~50% extra HG of modern Basques probably (rel. to EEF).
The evidence for this cline between types 2 and 3 is very apparent in ancient Basque DNA, as well as some scattered other clues from other regions. It's possible that, within the ill-studied type 3, which obviously expanded over the other two, there was some significant ANE scores (assuming ANE still makes any sense, because to me it does not).
@Simon: "I think ANE isn't that mysterious, it's an Upper Paleolithic genome from Siberia"...
I know what it is. The time distance between ANE and any European HGs is similar to that with farmers. ANE is way too old (from the LGM) and that may be why it is so problematic. Also ANE samples were all from Gravettian contexts, what is a Europe-originated culture (regardless of possible precursors in West Asia), so Ma1 and AG had EHG type admixture and that's probably what Günther & Valdiosera found in their repeated pre-Motala 40% admixture arrow towards Ma1. That is an additional confounding factor.
"ANE affinity hence may in principle come from various sources, but in Europe in general it means eastern admixture".
Or Nordic HG admixture.
@Maju: "Ajvide and the other Pitted Ware samples should have roots in Eastern European EHG-like populations, although maybe mixed with SHG type peoples."
Why do you think PWC people come from from the east rather than being, say, SHGs with a tiny bit of (very WHG-rich) EEF admixture?
@ Maju
It's possible that the ANE detected in SHGs diffused further westwards with Mesolithic cultures and that Loschbour wasn't typical for the more northern West European HGs. But even if this happened (not sure), the Northwest European HGs would in all likelihood have less ANE than the SHGs, at most the same amount, if complete replacement happened (unlikely).
Then, regarding a possible diffusion with SHG-admixed farmers: To my knowledge Gok2 has 0% ANE. It plots on the western end of the West Eurasian PCA, on a line between Sardinians and WHG, with no eastern shift at all.
I surely would be interested in seeing a WHG-EHG-EEF triangle model. I just doubt that it would be better than the K8 West Eurasian PCA, because the latter looks very similar to the genomic West Eurasian PCA which captures the basic pattern of West Eurasian variation. Thus the main factor of east-west differentiation is ANE affinity, with the ANE pole being at the imaginary meeting point of the northern HG cline and the southern West Asian cline.
I agree that the ATP samples are not the last word regarding Basque origins, and that slightly more northern samples would be even better, also from southwestern France.
BR1 from Gamba et al. is from the Mako culture. So far there is no autosomal DNA from Vucedol in the literature, but there is something in the pipeline, I've heard. But some Hungarian Bronze Age samples from Allentoft et al. are very similar to BR1, for instance RISE479 from the Vatya culture. In the PCAs he showed clearly an eastern shift and at the same time an unusually strong HG ancestry:
http://eurogenes.blogspot.com/2015/06/k8-results-for-selected-allentoft-et-al.html
@Simon: Lochsbour, La Braña and KO1 represent roughly the (epi-)Magdalenian genetic pool (at least as far as we know, a few more data points would be interesting to have). It is clear however that the areas around the North Sea were not quite just "Magdalenian" but something else (Hamburgian-Ahrensburgian, later Maglemøsean around Denmark) and that the Swedish samples are suggestive of important extra complexity that the basic WHG/Epi-Magdalenian pool can't explain on its own.
"I surely would be interested in seeing a WHG-EHG-EEF triangle model. I just doubt that it would be better than the K8 West Eurasian PCA".
It does not pretend to be "better": it just intends to quantify WHG vs EHG in a direct simple form in order to clarify this confusing issue. In principle everywhere where the Lazaridis' triangle detected ANE should produce correlated amounts of EHG but it should be something much more direct and hopefully whatever differences should be informative.
Complementarily EHG vs Yamna tests should help quantifying if there is actually some intrusion that is EHG but not Yamna-like. Another complement could be contrasting Ajvide or SHG vs EHG, in order to discern what is truly Eastern European of IE timeline and what is older. There's a lot of info that can be analyzed with current data along these lines.
"BR1 from Gamba et al. is from the Mako culture".
Thank you for the info. I will have to research this BA culture. In any case I am sure all of us would have expected at least some Yamna-like ancestry in it, yet nothing at all is apparent. For me it looks much like the Baalberge samples, which should have Yamna but do not (or is low in the best case).
Just checked and David's K8 does not have EHG (has ANE instead but not the same). I strongly believe that EHG would be much better reference, that it's time to ditch ANE and look at the real thing.
I'm quite convinced that there was also an ANE source radiating out of the Caucasus/Trans-Caucasus area with Kura-Araxes and later cultures. This is probably responsible for most of the ANE in the Near East, and from there it affected Southern Europe, most of all Greece, Italy and some of the islands. So we have three possible ANE sources: EHG, SHG and the Caucasus area. David's K8 indeed has an ANE pole, but no EHG (etc) components. But the PCA based on K8 data is still quite informative, as all the SHG, EHG and Caucasus/SC Asia samples are shifted towards the east, in line with their ANE affinity. Baalberge cannot really be compared to the Bronze Age Hungarian samples, because Baalberge has 0% ANE in the K8, and hence plots with other Middle Neolithic farmers on the far left side of the plot, whereas at least some of the Bronze Age Hungarians have a clear eastern shift.
@Simon: or again it is possible that the varied ANE degrees of affinity in West Asia represent an structure of the region in the Meso-Neolithic area.
Ma1 and AG are too old to be necessarily "recent" (Metal Ages') influences: they can be older. Also this Paleosiberian population was related by origins to West Asia, South Asia and Europe in at least a double way: (1) the Aurignacoid first layer (which is what arrived to America) and (2) the Gravettian overlay, which brought East European "paleo-genetics", which in turn may have some relation with some but not other West Asian populations of the time. We should not assume extreme simplicity but rather some complexity already in the Upper Paleolithic.
That's why I think that ANE should be discarded and EHG used instead (complementarily to Yamna).
Post a Comment