Are R1a and R1b really Indo-Europeans ?

[Kentel]

Here is a quote from a text I read on the eupedia webpage devoted to genetics :

The Indo-Europeans' bronze weapons and horses would have given them a tremendous advantage over the autochthonous inhabitants of Europe, namely the native haplogroup I (descendant of Cro-Magnon), and the early Neolithic herders and farmers (G2a, J2, E-V13 and T)

The theory according to which superior and patriarcal IE horsemen (Dumézil's mannerbund) arrived in Europe and subjected a population of inoffensive matriarcal pre-IE peasants whose languages they eradicated, has been supported from the very beginning of the IE studies, mostly on ideological grounds (see f.ex. Demoule http://www.anti-rev.org/textes/Demoule99a/ and http://mapageweb.umontreal.ca/tuitekj/cours/DeMouleMytheSurMesure.html).

This so-called invasionist model has moreover many flaws :

Archaeologic:

The debates are always been very vivid between linguists and archaeologists around the PIE question since archaeologists cannot really afford clear evidences to the conjectures of the linguists.

1- There is no objective archaeological attestation of a massive migration in Europe during the Bronze Age, that is : no archaeological culture spreading from the Black Sea to Western Europe at this date. The only continuous cultural horizon which can be connected with a migration in Europe is the dispersal of farming which predates the Chalcolithic (see Renfrew). Kossina's, Childe's and Gimbuta's theories are purely speculative since they rest upon jumps from a culture to another (eg. from Kurgan I,II,III to Yamna to Baden + Globular Amphora to Corded Ware II + Bell-Beaker to Unetice to Tumulus to Urnfield + Halstatt etc.)

2- The material found in the graves (Kurgans or others allegedly IE burial sites) showed no evidence of horse-riding.

3- All in all, it seems pointless to connect an archaeological culture to an ethnic group : technical innovation can spread by contact without any population move.

Genetic:

1- If I understand well R1a is mostly represented in Eastern Europe and R1b in Western Europe. We should thus have a linguistic split corresponding roughly to Slavic/Baltic/Hellenic/ on one side and Celtic on the other, with Italic and Germanic in the Middle. No linguistic branching agrees with such a split, as far as I know. It is even more problematic when we consider that this split between R1a and b occurred around 10 000 BC (Science 290) since it should have produced two clearly distinct linguistic branches at a very early date.

2- The fact that R1b is found at very high frequencies along the Atlantic coasts of Europe, in Siberia and in Central Africa is confusing, to say the least.

3- The dating of the genetic mutations is very inaccurate : f.ex. the G haplogroup has been dated from 30 000 BC (National Geographic Society), 17 000 BC (Semino 2000) and 9500 BC (Cinnioglu 2004).

4- 80% of the modern European genetic material dates back from the Palaeolithic according to Alinei.

5- R1a is connected with the Kurgan Culture, hence with PIE speaking people. I have no objection to this but what about R1b in the meantime ?

Linguistic :

Language variation is very slow, as far as we can observe it. There is f.ex. very little change between mycenean greek and modern greek (within a time span of 3500 years) or between old french and modern french (a time span of 1000 years), hence expecting a split of the PIE language families as late as the Chalcolithic seems unrealistic. In terms of language variation, the differenciation between hittite and latin would require much more than 1500 years.

To conclude this very long post (sorry for that), I have the feeling that linguists are trying to distort the facts in order to make them coincide with a highly ideological view of mankind (cf. Dumezil f.ex), based originaly on a poem from the Veda, namely the - very unclear - 96th Hymn to Indra.

 

[Yetos]

that is my question also,

I agree with arsenic bronze via tin bronze road, by I am not yet conviced 100% that R1b and R1a had IE as mother language and spread IE.

 

[Kentel]

that is my question also,

I agree with arsenic bronze via tin bronze road, by I am not yet conviced 100% that R1b and R1a had IE as mother language and spread IE.

Good to hear that I'm not alone, thanks

 

[Taranis]

First off, thanks for this elaborate and thought-provoking (and also a bit provocative at some parts, I appreciate that :grin: ) post!

It'll take a while for me to walk through all your points, because you brought up a number of very interesting ones, but for the moment, I will start with the issue of genetics and get to your linguistic points later:

2- The fact that R1b is found at very high frequencies along the Atlantic coasts of Europe, in Siberia and in Central Africa is confusing, to say the least.

R1b is dominated by three main subclades: M269 - in Europe, the Caucasus and Anatolia, M88 in the Near East and Africa, and M73 in Asia.

Western European R1b is part of M269, and from the looks of it, the result of a massive founder effect. Most R1b in Western Europe is part of the subclade L51, which is in turn dominated by the subclade L11. Conversely, L11 is basically nowhere found outside of Western Europe.

How R1b arrived in Western Europe (eg. maritime or land route) and from where (Balkans? Black Sea Area?) is still fuzzy as of the moment. What we do know is that R1b was absent in Neolithic sites in Catalonia and Germany (instead G2a was the dominant Y-Haplogroup there, which was also Ötzi's Haplogroup). The currently oldest find of R1b comes from a Beaker-Bell site in Germany. I'm not too happy with associating the spread of R1b with the Beaker-Bell Culture though, since the expansion patterns do not match: R1b-L51 makes more sense as having dispersed across Western Europe from approximately the upper Rhone area rather than central Portugal.

With regard for R1b-V88, as I argued in this thread, it is possible to associate the dominance of V88 and lactase persistence amongst the Chadic-speaking peoples (as well as their neighbours) with evidence for some migration from the eastern Mediterranean to subsaharan Africa.

3- The dating of the genetic mutations is very inaccurate : f.ex. the G haplogroup has been dated from 30 000 BC (National Geographic Society), 17 000 BC (Semino 2000) and 9500 BC (Cinnioglu 2004).

I'd like to hear what Sparkey has to say on the issue. I have to say this thought: at the moment we probably have no fully reliable way to disprove the hypothesis that Haplogroup G2 was actually in Europe since the Mesolithic or earlier, but on the flip side, I've not seen anybody argue for European G2 being Meso- or Paleolithic. I think, given the abundance of G2 in samples from Neolithic farmers, that G2 is most probable to be actually Neolithic.

4- 80% of the modern European genetic material dates back from the Palaeolithic according to Alinei.

Is that Mario Alinei? The same Alinei who posited the Paleolithic Continuity hypothesis? If yes, I'd like to know when he made that statement (I heavily suspect that it may no longer hold up).

5- R1a is connected with the Kurgan Culture, hence with PIE speaking people. I have no objection to this but what about R1b in the meantime ?

Well, we can also reverse the question: if we say R1b is the original Indo-European Haplogroup, what would this make of R1a?

(to be continued )

 

[kamani]

To my knowledge the oldest R1b in europe is about 2,800–2,000 BC, in Bell Beaker culture, Kromsdorf, Germany. This puts "R" in europe in
the Bronze Age, early Hittite empire timeframe. So far we don't go against mainstream IE theory.
However we have things like R1b1a2a L23/S141, time of origin 5000 BC, which is found in Greeks, Albanians, and anatolian branch of IE speakers.
Then we have things like R1b1a2a1a1, a child of the previous, time of origin 4000 BC, which is found in Italo-celtic and germanic IE speakers.
So an hypothesis is that R1b has been in europe before the bronze age, we just haven't found any ancient dna yet to prove it.

 

[Taranis]

Linguistic :

Language variation is very slow, as far as we can observe it. There is f.ex. very little change between mycenean greek and modern greek (within a time span of 3500 years) or between old french and modern french (a time span of 1000 years), hence expecting a split of the PIE language families as late as the Chalcolithic seems unrealistic. In terms of language variation, the differenciation between hittite and latin would require much more than 1500 years.

I think with regard for language evolution, these are indeed slower examples. On the flip side we have for instance Primitive Irish to Old Irish, or Old Brythonic to Old Welsh / Old Breton. I must admit that both examples are somewhat flawed in so far that Primitive Irish and Old Brythonic are poorly attested languages, but I think that they do hold up. But I agree nontheless, the thorn in the IE model are the Anatolian languages, which have both features lost elsewhere, and are lacking features found elsewhere. From this perspective, and "Indo-Hittite" model (that is, an early split between Proto-Anatolian, and the rest of the IE languages) is certainly interesting. Interesting point (which in my opinion is also a forceful argument against Renfew's Anatolian hypothesis), came from Melchert (2012), however: he rejects the idea that the Anatolian languages diverged significantly earlier from the rest of Indo-European (I'm ambiguous on the issue), but he brought up another counter-argument:

"I may add the further counterargument that, if the Anatolian Indo-European languages at the time of their attestation had been in situ for five thousand years, it is not remotely credible that they would show so few genuine loanwords from or into Sumerian, Hattic, or the nearby Semitic languages."


"I stress that I know of not a single compelling example for a pan-Anatolian loanword from any non-IE Near Eastern language."

In other words, if not even Proto-Anatolian was spoken in Anatolia, by what logic are we then to assume that PIE was spoken there? In my opinion, Indo-Hittite or not, PIE must have been spoken elsewhere. However, I would like to dwell with the Anatolian Hypothesis a while longer:

The main argument, from the linguistic perspective, for the traditional hypothesis is the vocabulary of PIE, namely common words for "horse" (*ek´wos), "wheel" (*kwekwelos) and "metal" (*Hejos). The combination of these three items, in my opinion, very much narrows down the context where and when PIE might have been spoken. (I must add though that "horse" and "metal" are by themselves weaker arguments: "horse" may not necessarily signifiy domesticated horses even if it seems likely, and "metal" may not necessarily signify knowledge of smelting/metallurgy.)

(I might also bring up warrior/military terminology, but I'm somewhat sceptical with this: how 'peaceful' were Neolithic societies really?)

The main counter-argument against the traditional reconstruction of PIE as a late Neolithic / Copper Age language would be that these could be wanderwörter (German for "wandering word"), the textbook example of which is the word "wine", found in the Indo-European, Kartvelic and Semitic languages.

A good example of another borrowed word, at a much later point, is the word "copper":

Greek "Kypros" (the island)
Italic/Romance: Latin "cuprum", French "cuivre", Spanish "cobre"
Germanic: English "copper", German "Kupfer", Swedish "koppar"
Celtic: Irish "copar", Welsh "copr"
Non-IE: Basque "kobrea", Finnish "kupari"

Note how when comparing Germanic vs. Romance, the word is ignorant of Grimm's Law (it obeys, however, to the Second Germanic Sound Shift exhibited by German), or how it's ignorant of the loss of *p in Celtic. And of course, the word is found in Non-Indo-European languages.

If Renfew's Anatolian hypothesis was correct, the earlier mentioned IE words (ie "horse", "wheel", "metal" would expected to have been invariably borrowed. But, how can a word be a 'wandering word' if it is attested at the geographically opposing ends (eg. Germanic vs. Greek vs. Indo-Aryan vs. Tocharian), and at the same time, be subject to the sound laws of the respective language branch? The consequence would be that an uniformous, unchanging form of IE was spoken over several thousand years from Northern Europe to the Tarim Basin, and I think that's a stretch of imagination.

With regard for Alinei's hypothesis (Mario Alinei), you have the same problems that exist with the Anatolian hypothesis, only multiplied by factor ten: we would have to explain why the Proto-Indo-Europeans had not only words for "wheel" and "metal", but also for "cow", "sheep", "plough" and "grain" if they had neither domesticated animals, nor agriculture.

There is, however, a reason for some clemency towards the Paleolithic Continuity hypothesis: I'm under the impression that Alinei mostly received attention based on the picture that genetics of the mid-2000s seemed to paint, namely that the genetic makeup of Europe had not changed significantly since the end of the last ice age... so why not argue the same for the linguistic makeup? However, the research results that were supposed to supported his hypothesis early on, are long-since outdated.

Which takes me back to the Anatolian Hypothesis: if we disregard items that were clearly only invented in the terminal Neolithic / early Bronze Age, Proto-Indo-European might be older... only, not in Anatolia.

 

[Kentel]

I had written a very developped answer to Taranis but have lost everything when sending the post I will have to take it back from the beginning, sorry for delaying my answer...

 

[Kentel]

Thank you Taranis for your very informative posts; it gives much to think about. I am not as comfortable with genetics as you are, thus some of my remarks may sound inappropriate, but I am really curious about the matter. Which does not exclude some pinches of provocation, indeed

Let's begin with the genetic issue : if I summarize the chronology in Europe, focusing on the areas with the highest frequencies:

1- Haplogroup I (map here http://en.wikipedia.org/wiki/File:Haplogroup_I.png)
- M253 or I1 in Scandinavia, North-Western Poland, Cotentin Peninsula in France, South-East England
- M223 in the Danube Delta, Northern Germany and in Brittany
- M423 in the Balkans, the Danube Delta and in the whole Romania, Bulgaria and Ukraine.
- M26 in Sardinia

point of origin : the Balkans maybe ?
date : around 25 000 BC ?

2- Haplogroup G
- P303 (G2a3b1) in Europe, especially in Greece, Sardinia, Spain and Central France.

point of origin : the Near-East
date : around 10 000 BC ?

3a- Haplogroup R1a (M420)

- M17 or 198 → M458 in Poland, Russia and Ukraine

3b-Haplogroup R1b (M343)

- P25 → P297 → M269 → L23→ R1b1a2a in Western Europe
- P25 → P297 → M73 → R1b1b1 in Siberia
- P25 → V88→ R1b1a → in Africa M88

point of origin : disputed
date : around 20 000 BC (R1a/R1b)


Is it correct ? I'd rather wait for your comments before drawing any conclusion from this.

 

[Yetos]

@ Kentel
R1a has quite enough diversity in Balkans,
it could be balkanic,
But until today that is considered as a sink phenomena, as area where R1a left marks due to invasions and devastations and not as primary homeland of R1a.

 

[Kentel]

@ Kentel
R1a has quite enough diversity in Balkans,
it could be balkanic,
But until today that is considered as a sink phenomena, as area where R1a left marks due to invasions and devastations and not as primary homeland of R1a.

Yes, in fact these data draw a picture of the actual frequencies of these haplogroups. I don't know if we have data from prehistoric samples, and if we have, how much. However, I guess that an area with a peak of frequency indicatesa very old presence (?).

The goal of all this is to redraw the migration routes of these populations, and afterwards to infere if they could have been IE people or not. Well, I and G certainly were not.

 

[sparkey]

I like these sorts of challenges. Very well structured, Kentel. My best crack at your enumerated points about genetics...

1- If I understand well R1a is mostly represented in Eastern Europe and R1b in Western Europe. We should thus have a linguistic split corresponding roughly to Slavic/Baltic/Hellenic/ on one side and Celtic on the other, with Italic and Germanic in the Middle. No linguistic branching agrees with such a split, as far as I know. It is even more problematic when we consider that this split between R1a and b occurred around 10 000 BC (Science 290) since it should have produced two clearly distinct linguistic branches at a very early date.

As Taranis points out, the type of R1b present in Europe is quite specific. European R1a is a bit older (actually it has quite high diversity in places like Russia) so it makes more sense to think of R1b as a more recent introduction. That doesn't mean that they both need to be certain branches of IE. A few patterns could explain modern distributions:

• R1b was a minority clade within the IE population at the foundation of IE, and not all R1b made it into that initial population (evidence: R1a and R1b correspond very well to IE distribution worldwide, better than almost all haplogroups, with R1a the better correspondence between them... see Iranian populations in particular)
• European R1b follows a founder effect pattern (possibly enhanced by cultural pressure)
• Satemization spread within a high-R1a subset of IE peoples, meaning a rough correspondence but not a perfect match between Satem and high-R1a

2- The fact that R1b is found at very high frequencies along the Atlantic coasts of Europe, in Siberia and in Central Africa is confusing, to say the least.

Different subclades, and it's always possible for expanding populations to accidentally expand haplogroups picked up from previous contact populations.

3- The dating of the genetic mutations is very inaccurate : f.ex. the G haplogroup has been dated from 30 000 BC (National Geographic Society), 17 000 BC (Semino 2000) and 9500 BC (Cinnioglu 2004).

It has gotten better. Nordtvedt's and Klyosov's methodologies are some of the best out there for STR dating, and others are beginning to look closely into SNP dating. Incidentally, Klyosov 2011's STR dating and Robb 2012's SNP dating calibrated to CT=70kybp converge to G having a 12,500 YBP TMRCA.

4- 80% of the modern European genetic material dates back from the Palaeolithic according to Alinei.

There is no reason that I know of to trust Alinei here. He just says:

[SIZE=-1](A) the areal distribution of genetic markers largely corresponds to that of the world languages (Cavalli Sforza et al. 1988, 1994, Menozzi et al. 1978 etc.);

(B) language differentiation must have proceeded step by step with the dispersal of humans (probably Homo sapiens sapiens) (idem).

(C) Independent geneticists working on DNA have recently ascertained that that 80% of the genetic stock of Europeans goes back to Paleolithic (e.g. Sykes 2001, 2006). [/SIZE]

(A) and (B) make sense, but (C) is just deferring to outdated Sykes conclusions. And interestingly, if we follow (A) and (B) but keep up with the latest conclusions relating to (C), we come to the opposite conclusion.

5- R1a is connected with the Kurgan Culture, hence with PIE speaking people. I have no objection to this but what about R1b in the meantime ?

Does this pattern work for you?: initial minority + amplification southward/eastward and diminishment northward/westward at first + minor drift into Eastern Europe after that + rapid founder expansion after that. That's one way I've thought of it, and it is probably one of many possibilities to explain modern patterns.

 

[LeBrok]

Yes, in fact these data draw a picture of the actual frequencies of these haplogroups. I don't know if we have data from prehistoric samples, and if we have, how much. However, I guess that an area with a peak of frequency indicatesa very old presence (?).

For old presence and origin look more at diversity of HP clades than at frequency. The more diversity you see in a region the more time it took to mutate and diversify, therefore it must have lived here for longer than in places of low diversity, and places where usually the upstream/younger clades are present.

Don't use Internet Explorer with Win7 to post, it's known to "eat" posts on Eupedia. Use Google Chrome instead, or other browsers.

 

[LeBrok]

Here is a quote from a text I read on the eupedia webpage devoted to genetics :


Linguistic :

Language variation is very slow, as far as we can observe it. There is f.ex. very little change between mycenean greek and modern greek (within a time span of 3500 years) or between old french and modern french (a time span of 1000 years), hence expecting a split of the PIE language families as late as the Chalcolithic seems unrealistic. In terms of language variation, the differenciation between hittite and latin would require much more than 1500 years.

I agree, languages evolve rather slowly without much of outside influence There are times though in history when this process is accelerated few folds, especially when one language is superimpose on other language substratum. Other words, language of conquerors over defeated population, or mixing peoples during big migrations, which we know happened few times in Europe.

We know fairly well what were the historical circumstance of Celtic Britannia evolving (Latin, Germanic, French) into English speaking Britain, both IE languages. The interesting thing is how grammar got simplified in the process Possibly due of locals learning a second language from minority of invaders, few times, butchering the grammar and pronunciation on the way. This is in span of 1,500 years.
Other example is how proper Latin evolved into pig Latin, and Roman Languages, when exposed as second language onto mostly celtic populations of European Mediterranean part, or just over other conquered Italian peninsula tribes.
If trend of grammar simplification is a rule of big mixing of peoples in the past, than we could use it to trace which EI populations didn't mix much preserving full or almost full grammar. It doesn't mean they didn't mix at all, but when they did, one language group was always overwhelming in numbers, which prevented grammar simplification.
It actually correlates nicely with Slavic expansion. The further they went south, the more grammar they lost. The simplest grammar in slavic languages is in Macedonian and Bulgarian, followed by Serbs I guess.

PS. I wonder if process is reversed when language is left alone for few thousand of years. Does grammar get complicated when tribe is insulated for that long? Why did EI get so complicated with 8 declensions, in first place?

 

[Kentel]

I agree, languages evolve rather slowly without much of outside influence There are times though in history when this process is accelerated few folds, especially when one language is superimpose on other language substratum. Other words, language of conquerors over defeated population, or mixing peoples during big migrations, which we know happened few times in Europe.

Exactly. The substrata question is the key. I have been working on this issue during the past 4 years - and I must confess that my hope with the genetic analysis (and our current discussion) is to be able to locate the pre-IE migrations and settlements in order to compare non-IE syntactic and lexical features within the so-defined areas.

And my conclusion is the same as yours : languages evolve rapidly when they come into contact with substrata. French is a Latin badly spoken by Gauls, as well as English is Old Saxon badly spoken by Brittons. And the reason why Greek is so stable is because the Greeks have not encountered any substratum during 3500 years.

I have discussed a little bit the subject on this forum a few months ago - and mentioned among other things the reluctance of the Academia and of the supporters of the Invasionist Model to accept the idea of substratic influence.

But this is another topic, and I will gladly open a new thread to discuss this matter

 

[Yetos]

@Kentel

what do you believe?
IE enter with R1b and R1a from steppe?
or existed in Balkans and minor Asia much before the R1b and R1a?


I mean considering that basquez is R1b population and Basquez language is Not IE, what do you believe? that Basquez is older or younger in Europe than IE?

could R1b or R1a learn IE language at the past outside steppe?? and then spread it?

 

[LeBrok]

I have discussed a little bit the subject on this forum a few months ago - and mentioned among other things the reluctance of the Academia and of the supporters of the Invasionist Model to accept the idea of substratic influence.

)

Not being a linguist, I had no slightest idea that this issue is controversial.
Hearing how Italians or Indians pronounce English, one can get an idea how sound shifts/laws could happen. I don't mean this is the only way, but we have so many real life and current examples to built on.

 

[LeBrok]

And the reason why Greek is so stable is because the Greeks have not encountered any substratum during 3500 years.

It makes sense, Greek was already a lingua franca before Rome expended to Greece. All educated Romans spoke Greek, so there was not much pressure to learn Latin. Later there was not much slavic incursion into Greece to influence any changes, except in Macedonia. And looks like Ottomans were quite tolerant and didn't force any community in europe to speak their language. Not many places like this in Europe.

 

[Dianatomia]

It makes sense, Greek was already a lingua franca before Rome expended to Greece. All educated Romans spoke Greek, so there was not much pressure to learn Latin. Later there was not much slavic incursion into Greece to influence any changes, except in Macedonia. And looks like Ottomans were quite tolerant and didn't force any community in europe to speak their language. Not many places like this in Europe.

How do you account for the fact that most of Anatolia spoke Greek in medieval times? Yet, during the fall of the Ottoman empire most of Anatolia spoke Turkish. What is notable though is that, after the fall of the Ottoman empire, Greek in Anatolia was only spoken in places where it was spoken prior to the Hellenization of that region. Meaning, the Aegean coast of Asia minor and some regions along the black sea. Makes you wonder how Hellenized Anatolia actually was.

I know that this is a bit off topic. But it puzzles me. Perhaps someone can enlighten me.

 

[kamani]

Language is notoriously bad at determining race or ethnicity. Entire populations can switch languages in a matter of a couple of centuries. For example, today there are close to 400 million people in south america, most of which speaks spanish. Based on how few spaniards actually moved to south america, most of these spanish speakers descend from something other than spain.
Sometimes there are even language revivals and resurections after the fall of empires.

 

[LeBrok]

How do you account for the fact that most of Anatolia spoke Greek in medieval times? Yet, during the fall of the Ottoman empire most of Anatolia spoke Turkish. What is notable though is that, after the fall of the Ottoman empire, Greek in Anatolia was only spoken in places where it was spoken prior to the Hellenization of that region. Meaning, the Aegean coast of Asia minor and some regions along the black sea. Makes you wonder how Hellenized Anatolia actually was.

I know that this is a bit off topic. But it puzzles me. Perhaps someone can enlighten me.

I was only referring to Europe in regards to Turks. I'm not too familiar with Anatolian situation much. It's probably safe to say that Anatolia was mainly Hellenized on sea shores, but not much in interior. It also might mean that Turks mainly settled in interior, and minorities of Greeks and Armenians survived fairly intact in their main centers by the sea. Well, at least till late phase when Turks became very nationalistic at end of 19th and beginning of 20th century.