Substack cometh, and lo it is good. (Pricing)


For some pieces on my Substack I’ve been re-reading a lot of the stuff on the ancient genetics and archaeology of Eurasia as they relate to Indo-Europeans. This means I get a different view from usual…as it’s more synoptic. I’m not entirely clear on the dates or archaeology, but here is what I’ve concluded: the Indo-European expansions can be partitioned into “waves.” That is, they weren’t a simple “demic diffusion” where disease (against their rivals) and reproductive excess generated a continuous expansion across their range.

So here’s what I get

1 – An “early phase” where Yamna people push west (Kurgan) and become Corded Ware, and east (far) and become Afanasievo. Date this to right before 3,000 BC, but pretty much “completes” in Europe by 2900-2800 BC, as the broad zone of Central and Northeast Europe is dominated by these people (there are still debates on whether Afanasievo became the “Tocharians”; I think they did)

2 – ~2500 BC, 400-500 years after the initial push west, Indo-European populations push beyond their limits on the Rhine, and breakthrough past the mountains ringing the Southern European peninsulas. The dates are often vague in the south, but it looks to be around 2500 to 2000 BC. For example, the Neolithic farmer descended Remedello Culture in northern Italy ends about 2400 BC. The Bell Beaker Indo-Europeans seem to have arrived in Ireland and England at just about this time, perhaps a century after they came to dominate France.

Though there were obviously islands of exception (often quite literally as in Sardinia and Crete), Europe by 2000 BC was Indo-European.

3 – The third wave dates to after 2000 BC, and it is the “Asia reflux.” Populations used the forest-steppe zone as a stepping stone out to the east. Derived from the same synthesis between Yamna and European farmer as Corded Ware, these populations seem ancestral to the Indo-Iranians. Slavic-speaking people (or the ancestors of those people) occupied the western fringe of this expansion zone, and by the Iron Age had begun to move east, marginalizing Indo-Iranians across much of their core European territory.

It seems that Indo-Iranians had pushed into the margins of northeast Iran, Khorasan, by ~2000 BC. In the period between 2000-1500 BC they clearly began to occupy their historical core zones in Iran and India. Obviously, Indo-European Iranians are present in western Iran by 1000 BC in the historical record, though Indo-European Mitanni are present by 1540 BC at the latest in Syria and northern Iraq.

The Iranians also moved into the Tarim basin, so the cities of the west and southern edge were Iranian-speaking (the cities of the north and east were Tocharian).

What explains these pulses? I don’t know totally, but we know a few things:

– There are star phylogenies on the Y chromosomal associated with these migrations. R1b, R1a, and I1. I think the last is due to the assimilation of non-Indo-European men in Europe, but the first two are clearly primal. The Indo-Europeans were clearly very patrilineal.

– The last, Asian, migration clearly has something to do with chariots and horses. The coincidence in timing seems too much. But the earlier migrations were before chariots (I believe). But, the horse does seem to have come with Indo-Europeans, so there was a level of mobility involved.

– The “Bell Beaker” motif seems to have emerged among non-Indo-Europeans in Iberia, and spread to Indo-Europeans, who expanded outward. I think we’re seeing something related to religion.

Unfortunately for I suspect that the Indo-European advantage was “social technology”, not material technology. Social technology is hard to infer in a preliterate society.

Question for readers: Can you nail down the chronology better? Those who know archaeology?

42 thoughts on “Indo-Europeans!

  1. Why didn’t the Y-chromosome groups of the Indo-european population that entered India displace the J2b2 male lineage if it was from IVC?

  2. why do you expect 100% displacement? that only happened in a very few places.

    going from a very small number to 20% or so seems like a lot

  3. Only wrinkle is military development:

    Wave 1: domesticated horses (still undersized), plague, wagons; I wonder if the use of youths as warriors (koryos) across IE languages is because only younger people could ride the initially smaller horses.

    Wave 2/3: chariots + bronze-working (the archaeological evidence for chariots in 2 is meagre, but Celts/Romans have chariots continue to play a symbolic role well after they’re no longer useful militarily, implying they played a traditional role.)

    Now it’s IE-speakers versus IE-speakers, but the Scythians/Iranians then breed horses large enough for adults & armored adults to ride & fire arrows from, leading to the obsolescence of chariots. I had always thought this was an east-to-west movement because the Sarmatians have far higher ANE admixture than the Srubnaya/CWC. We also see an increase of East Asian ancestry on the steppe.

    Then the stirrup drives another east to west wave. I have a very hard time understanding the Slavic expansion.

  4. Hmm… Ok, I’ll go a little full on geek out here and write a massive essay, on just one aspect I’ll get fixated on that I’ve looked at in detail. Sequence of entry into Europe (or further west into Europe if people prefer) of “steppe ancestry”. I’d say as present what you’ve put seems broadly correct to me but I would stress that it looks more in adna like a slightly later, then continuous and rapid process, rather than marked by a strong hiatus at the Rhine. That is, I don’t see it is as “quick regional takeover and admixture up to the Rhine, followed by a 400-500 year hiatus, then another rapid breakthrough” (hope I’m not misrepresenting anything).

    When I look at the samples we have, using Eurogenes G25 and the date information from Human Origins, the sequence seems like:

    1) The first Corded Ware samples showing up in the Baltic (Plinkaigalis242 Lithuania and I4629 Latvia), with very little or no admixture, into territory that seems to have been unoccupied by EEF farmers (or at any rate there are no samples of such), so not really displacing anyone else much. These are dated at 2850 BCE (about 90% Yamnaya ancestry and the rest is HG, which may or may not have been acquired locally at all).

    2) Then we next have a Corded Ware sample in Central Poland (poz81) at 2700 BCE, again not much EEF (93% Steppe, 6% HG), so nothing much indicating that ancestors drawn.

    3) But in very quick succession, the first steppe ancestry samples appear in Switzerland at 2675 BCE (MX304) and 2635 BCE (Aesch25), with 43% and 76% Steppe ancestry. These folks are not clearly from a Corded Ware burial, and we see a sample at 2500 BCE in France (CBV95, Northern France) with 76% steppe ancestry. Switzerland is fairly interesting here because there is a good sequence, unlike other regions like Southern Britain where you have cremation or acidic soil in Netherlands. There are still some samples in Switzerland by 2300 BCE and 2060 BCE who seem to have no steppe ancestry (samples MX193 and SX17, SX21, SX26). Corded Ware burials do show up in Switzerland at 2500 BCE (the males unusually had y-dna I2).

    I believe the earliest steppe ancestry in Sweden is also at a megalithic burial site (sample oll007 dated 2620 BCE), with the last preceding EEF sample at 2705 BCE (another megalithic site, sample ans016). Earliest steppe ancestry samples in Czechia 2575 BCE, with about 55% EEF. (Hungary is another region of interest, but not well covered, no samples between latest EEF from Baden Culture at around 3000 BCE, then first Steppe ancestry samples, Bell Beakers, around 2350 BCE).

    4) Add to this in Poland, there are some Globular Amphora samples all the way up to around 2700 BCE with 0% steppe ancestry. Also Globular Amphora samples with 100% EEF are in Ukraine as late as 2730 BCE, after the earliest dated Yamnaya at 2950 / 2850 BCE. The archaeology might be different but from the genetic side of things, the samples don’t seem to speak too much to a well-established Corded Ware dominance of Central Europe as of 2700 BCE.

    To me it looks perhaps a little more like Steppe ancestry wasn’t really anywhere in North Europe west of the Baltic around 2700 BCE, as far as the samples we have say (further may change the picture), then was very quickly everywhere as far as Switzerland, (and possibly beyond) to at least some degree by 2600 BCE. It doesn’t look like a two-step Europe thing, but just a one-step that happens explosively around 2700 BCE and then continues more or less constantly until “complete” after 2000 BCE. These dates are not totally precise (ranges), so it could be more like actually the steppe ancestry is earlier in Poland and later in Switzerland, but it seems less likely.

    I don’t think things really properly homogenize until at least 2000 BCE, as in Germany you still have people with 90% steppe at 2295 BCE (sample Corded Ware I1538, not a significant outlier to other CWC DEU at the same time), whilst there are Bell Beaker people in Bavaria with as low as 20% Steppe ancestry (sample I3594 dated 2175 BCE). This may relate to continued flows from outside the region though.

    SE Europe is probably a slightly different story, but there are fewer samples and I’ve looked at this in less detail. (But there seem to be a lot of Yamnaya samples coming from Hungary and other places that can be talked about in this regard, although they are possibly not very ancestral to later cultures.)

    Here’s this displayed with a quick graphic that I’ve posted before on Open Thread: . Also just to add that above “Steppe” is “Relative to Yamnaya Samara”, who probably have some low level of EEF ancestry).

  5. I’m not convinced the Bell Beakers were 100% Indo-European speaking. If they were, how do we explain the Basque being virtual carbon copies of ancient Iberian Bell Beaker samples (and 80% R1b-M269)? Either at some point the ancestors of modern Basques were assimilated into Vasconic culture without much of a genetic impact, or Bell Beakers (and maybe the preceding steppe cultures) were multilingual.

    And I’d note that tribal confederations on the steppe in the historic record were often multilingual, so I think one has to at least be open to the idea that some of the prehistoric tribal confederations that these cultures represent may have been multilingual and multiethnic too.

  6. If the Indo-Iranians expanded into Iran and India between 2000-1500 BC, then why is it Indo-Aryans who establish Mitanni in 1600 BC ? What route did they take ? I think the Indo-Aryans expanded first from Central Asia simultaneously into India, Syria and the Tarim basin between 2000-1500 BC. The Iranians expanded later into Iran proper, after 1500 BC.

  7. matt, do know you know what archaeology says?

    If the Indo-Iranians expanded into Iran and India between 2000-1500 BC, then why is it Indo-Aryans who establish Mitanni in 1600 BC ? What route did they take ? I think the Indo-Aryans expanded first from Central Asia simultaneously into India, Syria and the Tarim basin between 2000-1500 BC. The Iranians expanded later into Iran proper, after 1500 BC.

    that is a standard model. doubt genetics can answer since the two groups are so similar (same people?).

    some linguistics tho argue mitanni aren’t indo-aryan, but BEFORE the clean split btwn the two.

    i have no idea/don’t really care

  8. Razib: I listened to the podcast you did on the Sumerians. Here is an article that argues the Ur Kasdim of the Bible is actually modern-day Urfa in southeastern Turkey, 44 km north of Ḥarran.

    Ur Kasdim: Where Is Abraham’s Birthplace?–

  9. Some broad additions and questions I can think of, excuse the verbosity:

    There’s the apparent earlier migration that seems separated enough to be called a “wave” (it had been characterized in the literature by Gimbutas, Mallory, Anthony and others though I’d have to re-read the relevant parts to mention specific details; though it’s probably known to readers interested in the topic) that has given us a couple of earlier Chalcolithic outliers in the Balkans with steppe ancestry so far, one with R1b and one with G2a (there was another supposedly Varna individual that’s apparently problematic in its assignment and does look very modern overall), and is hard to know what happened to it, whether it assimilated or moved on, but it’s plausible to connect it to proto-Anatolian. We also have a relatively early low-quality individual from Kumtepe in northwest Anatolia around a period of change for the area that might have steppe-related ancestry but it’s hard to tell with absolute certainty. Dated in the paper around a millennium after the non-steppe admixed individuals at any rate, at ~5500-4900 BP.

    Potentially somewhat related to this early “wave”, there’s a Boleraz sample from Hungary (I2788) dated ~mid 4th millennium BC that also looks like it _might_ have small amounts of steppe-related ancestry since it’s prominently shifted compared to the others towards steppe-rich populations. Also some kind of G2a.

    A lingering question is the exact relationship between Corded Ware and Beaker considering the different lineages that characterize them, with the exception of some specific and sometimes later/less clear contexts so far. Same population with different founder effects happening at certain locations or different migration waves of related populations to the west, relatively early on? This would impinge on the question of separate “waves” in particular I think. Reference here also to the early western not-obviously CW-related R1b samples that Matt mentioned. This area needs further clarification but David @ Eurogenes has brought up some R1b-rich CW-related future populations before that will be published later on apparently.

    The steppe-related admixture that also seems to arrive somewhat earlier in parts of the Balkans than the rest of southern Europe (which isn’t particularly surprising considering the Danubian route) seems also somewhat uncertainly connected to the “northern” wave, if we’re being more nitpicky about it, and how exactly they separate from each other. There’s also greater apparent continuity of non-steppe lineages so far, compared to the significant Y-DNA turnover in Beaker-dominated (south)western Europe so the process seems potentially different. The appearance of steppe ancestry as far south as northern Greece, right on the northern border of Thessaly, happens by the late 3rd millennium – early 2nd millennium judging by two early Middle Helladic samples whose related paper isn’t out yet.

    This fits in well with the archaeological destruction phase towards the end of the Early Helladic, relative cultural impoverishment of the MH and subsequent rebuilding during the LH/Mycenaean. Then ~2000 BC sounds about right for the vast majority of Europe since we have samples from basically everywhere now. [Though to be all-encompassing, I’ll note that he relevant horizon for that isn’t completely certain. It looks like a Yamnaya-Balkan mix to me of some kind but will need more samples to clarify, whether to accept or reject instead an alternative scenario where it comes from that CW-related forest steppe reflux to the east that also seems to give rise to Indo-Iranian, coming back to the Balkans with the chariot]

    Also we have some recent samples from northeastern Romania with R1a-Z93 that are either dated relatively early (S11955/I11955, assigned 3500-3000 BC though from a 1958 paper) or have wide, potentially early dating (I11910-11-12 and I11913-14-15, assigned 5,450-3,050 BP) and look like steppe-farmer-ANE/ENA mixes of some kind. Whether their eastern kind of ancestry is more of the ANE-rich/Steppe Maykop kind or the ANE/ENA-mixed Cimmerian-Scythian-Sarmatian kind might help with their potential dating. We discussed them a bit in a previous open thread with Matt though I’m not sure if he has taken a particular look at them himself in Global25 yet. They might not represent something major in the IEzation of Europe either way but they’re interesting since at least one is potentially quite early and _if so_ might represent admixture between CW’s direct ancestors, Balkan farmers and Steppe Maykop-like populations.


    It might be, especially if we consider those later steppe cultures as representing proto-Aryan in the east and proto-Iranic in the west (e.g. Srubnaya) rather than proto-Indo-Iranian as a whole deriving from the eastern set (and that apart, some of the lingustic relationships within Indo-Iranian seems a bit complicated from what I read). It’s possible early West Iranic started spreading into its home territory only as late as the LBA-EIA of the region, as you said. Maybe the methods Reich et al. are looking into applying to ancient samples will clarify the more specific relationships which would be interesting for the finer details.


    With the current post-Beaker samples we have, Basques seem to have somewhat more steppe-related ancestry than the northern Iberian BA set and the southern (though also very limited) French BA set and appear most similar to the Iberian IA set. This might be an issue of not capturing all the variety around the area, since it’s not sampled _that_ well so far, but it might be the case that Vasconic speakers kept receiving decent amounts of admixture from more steppe-admixed neighbors (presumably some IE-speaking too!) even later on without linguistic change.

    It’s not necessarily as dramatic as the Beaker transition (though if these later groups are already similar to each other, due to more similar levels of steppe-EEF-HG ancestry, the turnover can still be great despite smaller apparent impact) but at least it opens us up to the possibility that pre-Beaker languages were simply preserved in some locations despite admixture with steppe groups and significant Y-DNA turnover at some point.

    We also have few Etruscan samples to make a really robust argument about their case either way but aside from a very north African admixed outlier, one is Italic-like and the other Italic-like with some apparent Balkan admixture, from the other side of the Adriatic i.e. basically like their IE-speaking neighbors. It’s not clear that Etruscan and Basque are necessarily recently related by the time they’re known so that might add another issue as to how many different potential families were brought to their western location over from the steppe, if we follow that kind of scenario, not that we should completely discard it out of hand either. I agree that the Basque (or Etruscan) kind of situation is far from satisfyingly clarified so far either way, even if we presuppose Chalcolithic pre-IE SW European descent for both and that the local Beakers went “native” over time.

  10. Though regarding one specific point, I’m guessing Reich et al. are already looking into recent connections between Beaker and Corded populations, like that Yamnaya-Corded sneak peek…

  11. I know you said it’s somewhat dated but the bibliography for Anthony’s “The Horse, The Wheel and Language” is full of references to Russian archaeological work on PIE and IE steppe peoples and their equine technologies.

    Is language barrier the only reason not to draw on those sources? They might suggest more precise dates for when and why the pulses occurred.

  12. Just why exactly did Iranians expand so much everywhere like Mongolia,Ordos Plateau,Xinjiang,Iranian Plateau and East Europe but Indo-Aryans did not?

    I find that the Academia seems to be very lenient in assigning Iranian Identity to every lesser known group like Scythians,Yuezhi,Wusun and all others but seem very reluctant to assign any Indo-Aryan identity to even the clearly Indo-Aryans groups like Mittani.

    Is it a kind of bias against Indo-Aryans in Academia?

  13. @Forgetful

    It is possible that in the case of Basques, something similar to what happened in Melanesia occurred. We now know from the Lapita ancient genomes from Melanesia that they were almost totally devoid of admixture from Papuan-like Melanesian natives, but as time went on Austronesian languages spread among Melanesian natives as male Melanesian natives mixed with the females of the Austronesian-speaking Lapita colonist remnants while replacing their males and acquired the language of their wives during the process facilitated by the Lapita matrilocality. So as a result, Melanesians today are heavily Papuan-like in Y-DNA and autosomes but with high Austronesian ancestry in mtDNA despite speaking Austronesian languages today. Something similar may have occurred in the Basque case, steppe-originated R1b males with Indo-European language might have mixed with EEF females in SW Europe speaking the precursor language to Basque while replacing their males and acquired the language of their wives. I read that Basques were traditionally matrilocal, if so, it is entirely plausible that Basque spread among Indo-European colonists the way I describe.

  14. I have seen the tendency in Western Academia to assign Iranic Identities to Indo-Iranian tribes with uncertain affiliation,very easily while Indo-Aryan identity is often questioned or rejected even if the evidence is reasonable.

    Mittani have clear Indo-Aryan Gods,Words and Way of life yet they are not Indo-Aryan.

    Kassite elites have Gods which seem to be Indic yet still they are not Indo-Aryans.

    On the other hand,we have but little attestation of Scythian language but they are unequivocally considered Iranian.

    We have so many Steppe Tribes all over Eurasia but all are declared Scythian or other Iranics.

    Yuezhi and Wusun are so surely declared Iranian.

    Ordos Culture,Pazyryk Culture and even Maeotians are unequivocally declared as Iranian.

    What is this academic indifference that Indo-Aryan archaelogical trail and Cultures is largely neglected?

  15. @Razib, no, sorry unfortunately I’ve only got Wikipundit level knowledge of the Mitanni, and no strong ideas about it. I don’t know how you’d pick apart the Indo-Iranian groups into separate waves really either; the linguistic and genetic and mythological divergences would probably have been pretty shallow at this time (and might not clearly reconcile with the Mitanni Indo-Aryan names on one side or other of the split)?

    @CupOfSoup, I think “the steppe” could have maintained multiple languages, if we consider that there was probably some group diversity there… But if the genetics do end up pointing to Beaker, Corded Ware, Yamnaya, Afanasievo all being somewhat separate patrilineal clan expansions from single population of perhaps a few thousand people, at around at least 3500 BCE, living in a small area, with many shared fourth, fifth cousins (*if*, this is unpublished), then it does seem hard to think it’s likely that the steppe ancestors of these groups spoke very diverged languages before some of them departed the steppe. (There are imaginable scenarios where one or more of these groups could’ve switched languages without much geneflow, but it is relatively less probable?).

    On the other hand, the opposing argument seems to be that because Basques are broadly similar to Bronze Age Iberians (difference noted above by Forgetful), and language expansions often patrilineal, then there must have been a very large proto-Vasconic sphere despite this leaving really no direct linguistic evidence at all outside of Basque people today (and very debatably Iberian inscriptions that are not agreed to be related to Basque either at all or in any kind of recent timeframe that matches a 2500 BCE expansion). I’m probably describing this unfairly and others can give more favourable spin (there’s also a feature some like where this would explain a supposed Basque-Caucasian language relationship). But that’s the jist as I see it. It just seems easier to imagine that there was some bilinguality in some communities and then changing of languages for reasons we don’t really understand. I suspect that people sometimes change and adopt languages for reasons of group differentiation and it may have made sense to emphasise a different language from neighbours at some point, and certain myths and legends.

  16. The spatiotemporal spread of human migrations during the European Holocene

    “Regardless of the model used, we found that HG [hunter-gatherers] ancestry is positively associated with broad-leaf forest anomalies, but negatively associated with arable land anomalies. YAM [Yamnaya steppe peoples] ancestry, in turn, is positively associated with pasture/natural grassland..

    We did not find that NEOL {Neolithic farmers who migrated from the Near East] ancestry had a strong association to changes in vegetation. One possible explanation is that this association was too minor or localized for us to clearly detect an effect in our model.”

    You also have:
    Megadrought and Collapse: From Early Agriculture to Angkor

    Which more directly addresses population movement/change due to climate shifts.

    A more modern example of pastoralists pushing into agricultural areas can be seen in

    Cotton, Climate, and Camels in Early Islamic Iran: A Moment in World History Paperback – July 5, 2011

    Given that horse-based people tended to have military advantages, it seems likely (not very surprisingly) they would be able to push their way into farmland territories that were under stress from climatic change. With sufficient momentum, they might be able to just keep pushing.

    My point would be that the technological advantage was there with the YAM, but not sufficient by itself, to cause the population shifts. The NEOL appeared to have a greater technological advantage over the HG. But even here, it required a stable enough climate to make farming possible.

  17. One theory for “why” I haven’t seen too much:

    1) Starts out like David Reich suggests – wagon use on the steppe expands range of pasturage, more domestic animals, which causes population boom.

    2) Then the amplifier is the social advantages of being from a recent population boom push further expansion. I’d guess for groups with a similar agro-pastoral base, the ones where more people who have lots of recent shared language, kinship, myths, male a shared recent male common ancestor, probably would have an advantage in struggles for land and pasture. People who are having a population explosion might also have more young people. The advantage is momentum? In Europe at this time there are lots of separate material cultures, so possibly a lot of separate languages and identities and such.

    Seems more like something we can test I guess. (Whereas ideology type stuff is… You can only know about IE through cultural reconstruction of myths, words and can’t really know about pre-IE at all very well.)

  18. It now seems like the biggest remaining mystery regarding the origins of Indo-Europeans has to do with the Anatolian branch (due to our lack of ancient DNA from the Hittites). Basically, are they descended from Yamnaya as well, or are they the descendants of a migration which predates the Yamnaya.

    IIRC, Reich is leaning towards the original (pre-Anatolian) homeland of the Proto-Indo-Europeans not being in the steppe, but somewhere south of the Caucasus, right? This would imply the Hittites were not particularly intrusive to the area, and will not show steppe ancestry.

    I seem to recall there are some Hittite remains now under investigation for ancient DNA, so hopefully this will be answered soon. Although given the lower classes spoke Hattic among themselves, it might be some time before we can sequence a “real Hittite.”

  19. @Razib @Forgetful, @Matt – Basque being the remains of a pre-steppe Beaker language doesn’t explain why their Y-DNA is overwhelming from the steppe-derived Beakers though.

    I just have a hard time believing one can have 80% of their Y-DNA pool replaced without a language switch. Especially in what seems to be very patriarchal cultures (ie the steppe cultures, not the Basques themselves).

    I agree the initial expansion probably had a single source, but I could see more diverse groups accreting on to the IE expansion as it grew, and given the diverse roots of the original steppe culture, linguistic diversity at least seems possible.

    Consider that Yamnaya is a compound of CHG, EHG and EEF, and that EHG is itself a compound of WHG and ANE groups. That’s 5 deeply diverged groups contributing ancestry. Even if we only consider the Y-chromosome pool origins seem to be divided between WHG-related lineages (R1b, I) and ANE-related lineages (R1a, Q).

    I look too at my own part of the world (British Columbia) and there are literally 30 languages indigenous to this province. Off the top of my head (apologies to any groups I forget), the major language families are linguistic area is centered on the Salishan, Wakashan and Chimakuan, Tsimshianic, Chinookant, Sahaptian, Kutenai, Na-Dene, Haida and Algic. There are no convincing links between these language families other than areal influence on one another.

    That’s in an area about 1,000,000 square kilometers in size – roughly the same as the Pontic-Caspian steppe – and without particularly sharp genetic divides.

    Granted, mountains create a lot of barriers here, but 10 families leaves plenty of room for 2 on the steppe, no?

  20. @Razib @Forgetful, @Matt – Basque being the remains of a pre-steppe Beaker language doesn’t explain why their Y-DNA is overwhelming from the steppe-derived Beakers though.

    basque were matrilineal

  21. “basque were matrilineal”

    I guess the question is, why did basque remain matrilineal when there is 80% replacement of males with a strong patrilineal culture? Wouldn’t that disrupt the original matrilineal culture and subsequently language?

    Seems inconsistent if 20% male presence was enough for cultural dominance elsewhere.

  22. there is zero evidence it happened in one generation. so every generation a minority of males were outsiders and eventually their children assimilated. over enough generations you can have population turnover if you have a smaller group surrounded by a larger one

  23. “so every generation a minority of males were outsiders and eventually their children assimilated. ”

    I guess the question is still about why these children are being assimilated instead of males asserting their patrilocal/patrilineal culture. Particularly when new waves of males are arriving with each generation.

    Not that we can figure out this precisely, but it still appears anomalous in larger context.

  24. @CupofSoup: “I just have a hard time believing one can have 80% of their Y-DNA pool replaced without a language switch.”

    So, how do you explain Vanuatu?

  25. A part of the basque case could also be the R1b input lineages simply being more likely to have Male children. Like a somewhat higher proportion of y chromosome sperm cells. Not sure about the current consensus regarding this idea.

  26. DaThang, not crazy but it seems kind of hard to credit really on what we can observe – no other male shift with other R1b like R1b-V88 bounce back in Europe generally (or much later, and only in Sardinia), a specific R1a subclade expansion at same time (while general parents of each have nothing special happen), introduced in a big wave with autosomal ancestry, no evidence of this happening in present day people. Crucially, you’d also see persistently skewed sex ratios in R1bized populations across history, lots more men with no wives, excess births of boys, which we don’t. It seems like even a sustained 1% increase per generation would only predict slow change (not rapid replacement), but have by now driven other haplos to low freqs wherever introduced (not the case).

    (In general my prior is that any hypothesis linking any functional trait to big y-haplos is fake, until and unless it’s strongly evidenced in Biobank scaled datasets…. Which nothing been ever has in this direction as far as I can tell, for any trait, even sex linked ones. There might be some subtle traits out there that would be cool, but it doesn’t seem like anything large or major has yet jumped out. The null hypothesis seems to me that high level haplogroup defining y mutation variation in humans contributes little to phenotype.)

  27. “what is anomolous. it’s a matrilineal culture. how do you think those operate?”

    A few patrilineal culture males enter a matrilineal culture. When they take brides, do each of them go to their wife’s household or create their own clan group?

    You seem to suggest that it is perfectly reasonable that they should each enter their wife’s household/clan since wife to matrilineal.

    I think it is anomalous because they are patrilineal and would prefer to live with their own clan of brothers (where males are related) rather each of them form new bond with their wife’s brothers (or other male groups provided by their wife’s family).

    Also, they would prefer to keep their sons with them because of their cultural preference of keeping father’s name intact. This changes the matrilineal culture to patrilineal over time where each new generation of new-comers will take more brides and keep with their clans rather than their wife’s family. Theoretically, this should change the language too.

    If they can so easily assimilate to a matrilineal culture, it points to them not having a strong patrilineal culture themselves. This again seems odd given how they imposed strong patrilineal cultures wherever they showed up (i.e, keeping lineage lists).

  28. One quick thing to add about how models of adna seem to complement linguistics.

    One linguistic finding is that IE seems to have a large number of primary branches, with no well agreed upon higher structure to the tree. That is, the IE languages that have been attested fall easily into the main subgroups (Celtic, Germanic, Armenian, Greek etc), but unifying subgroups together is very hard and there is no real consensus. This is more so when items like the satem change or Ruki change, which are noted to be contact driven between separately evolving branches, are taken into account.

    (Quote: Garrett, 1999, a “question posed in recent work by Johanna Nichols. Like many profound questions, this one is both shockingly obvious and disturbingly obscure: Why does Indo–European have so many branches? … Attempts to construct a highly articulated Indo–European family tree beg one crucial question: Why are well–established subgroups like Greek and Indo–Iranian defined by many distinctive innovations, while higher–order groups” (which are never raised without dispute) “are defined by only a few? … According to Nichols, the explanation has to do with the dispersal of Indo–European. She argues that ‘[m]ultiple branching at or near the root of a [family] tree points to abrupt dispersal of the protolanguage in a large spread’ (Nichols 1997b: 371). A ‘spread zone’ is defined as ‘an area of low [linguistic] density where a single language … occupies a large range, and where diversity … is reduced by language shift and language spreading.”)

    That suggests a “starburst” expansion of IE, followed by some contact, then later mergers, and then further expansion of some local varieties (in Iron Age perhaps).

    That fits with a model where expansion to Europe and the steppe lacks any strong “stage” structure. If we had, for instance, a Indo-Balto-Slavic stage evolving together for hundreds of years, or a thousand years (1/3 of time since split from pIE), we might expect to see a shared strong linguistic signal defining a subgroup. Whereas typically a signal is more like a flat “rake” structure, with some evidence of contact between European branches (Italic, Celtic, Germanic, Balto-Slavic), but which still doesn’t resolve clearly as a family.

    That seems to fit well with the adna, where even groups sharing a patriline (like R1a between Indo-Aryan and Balto-Slavic) don’t really have it shared much after expansion of IE at 3000-2800 BCE, or even a split in patriline before that time before, and also where adna doesn’t seem (to me anyway) to indicate much structure in expansions, just a quick starburst.

    In a staged expansion, we’d see high easily resolvable structure in family grouping, where very little is easily resolvable and only some dialect contact, which seems more “star-like”. So to me that seems to speak to rapid single expansion being more important in IE, and that fits with a rapid, relatively stageless chronology for Europe.

  29. If they can so easily assimilate to a matrilineal culture, it points to them not having a strong patrilineal culture themselves. This again seems odd given how they imposed strong patrilineal cultures wherever they showed up (i.e, keeping lineage lists).

    i’m trying to be nicer to people online. so it seems you are a person of low acuity. if they choose to move to the territory of the basques with a very different culture you can’t assume they are a random sample of the source population.

    in other words, they are going to be men who are somewhat different in their cultural origins from the source (e.g., they may be younger sons who lack inheritances or something like that, explaining why they leave).

  30. @Razib – “they may be younger sons who lack inheritances or something like that, explaining why they leave”

    That’s an interesting thought. If clan heads were practicing mass polygamy (which seems likely) then there would be a natural surplus of unattached males going out looking for prospects among other groups and cultures.


  31. @Matt can you point me to the Garrett work you’ve quoted here? Looks interesting, thanks!

  32. @CupOfSoup, there could be polygamy, though as far as we can tell so far, at least the Beaker people actually *seem* monogamous from recovered samples lacking any individuals who are related as brother/sisters through fathers, but not through mothers, even in quite large sets where there seems to be patrilocality. (Only example of a polygamous family indicated so far in this way seems to have been found in a Globular Amphora massacre in Poland). Although it is relatively rare to find relatives.

    However there could have been some other mechanisms creating male excess, like providing preferential nutrition and so on to sons (as children and as adults), and then status hierarchies in who has access to women within the group. And this would create similar incentives. Such things needn’t be vastly different to EEF societies, which were probably mostly patrilineal, patrilocal and preferentially provided access to males (at least as far as some limited data I can remember on nutrition in samples from Spain) but different enough on average to push the incentives for Indo-European groups to form more bands of males looking for females from other vulnerable local societies.

  33. About the initial move out of the steppe, the Holocene Optimum period seems to end about that time, so I would guess colder climate drove herders to new, warmer land. At the same time, Neolithic farmers were collectively drove South by the same phenomenon, as they died off on their traditional lands, moved themselves South or were push by incoming herders (probably, all of the above). At that time, having livestock as their main source of food was a better use of the land than crops, which made them more successful in the region.

    About the Bell Beakers, when these herders reached to Iberic Peninsula, they met a prosperous proto-civilization. They found themselves a niche as herders/traders/mercenaries and did the same as Normans millennia later, they adopted the language and social order in which they meshed in. In turn they would expend the trade network of early Bell Beakers much farther North and East that they could have done with their Neolithic farmer culture. The herders would also find new lands for their livestock in the mountains; to this day, the Pyrenees agriculture is centered around livestock while the valleys bellow are more crops intensive. The same would be true for other mountain ranges in Europe and lands farther North were crops would fail more often.

    As for the Basque, we need to remember that if they are an isolated language nowadays, there weren’t always so. Just back to Julius Caesar’s days, the Aquitanian and Iberian language families would make proto-Basque part of a continuum of agglutinative languages.

  34. @Matt, @CupofSoup,

    I came here to posit the same solution of polygamy of how a matrilineal culture can continue with arrival of a set of patrilineal males.

    If a single male can roam between multiple female households, then matrilineal culture would hold while leading to a change in Y-haplogroups.

    Contrary to general Bell Beaker Culture, only among Basque is there a lower variation in possible Y-haplogroups compared to their surrounding cultures?

  35. @Razib,

    “i’m trying to be nicer to people online. so it seems you are a person of low acuity.”

    We have different ways of reasoning from the same set of information. Sorry to have caused you impatience by trying to engage in a discussion on your forums.

Comments are closed.