August 26, 2009

Bronze Age origin of Semitic languages

Bayesian phylogenetic methods, originally developed for biology, have been increasingly -and successfully- applied to linguistic data in recent years (e.g., for Indo-Europeans, Melanesians, and Austronesian speakers from the Pacific).

The current paper proposes a Bronze Age origin for Semitic languages, ~3 thousand years after the split of European from Anatolian Indo-European speakers. I don't find this particularly surprising, as Semitic has been, until relatively recently, much more geographically constrained than Indo-European, and -due to the early literacy of the populations of the Near East, its post-Neolithic arrival can be observed in the archaeological record itself.

It also explains a facet of Y-chromosome distribution, that I have commented on before, namely the fact that the common Near Eastern haplogroup J2 extends from Europe to South Asia in a "horizontal zone" accompanied with little of its sister clade J1, but in the Near East itself, there is a "vertical zone" from the Black and Caspian seas to Arabia of high J1 frequency. As I have explained recently, the mixed J2/J1 frequency in the central Near East is due to an enrichment with J1 lineages of a population that had (in pre-Semitic times) a high J2/J1 ratio like those of Europe, Asia Minor, and Iran. J1 should not be seen as exclusively Semitic, but it can't be denied that the major factor affecting its current spread has been the arrival of Semites from the South, the latest episode of which involved the spread of Arab Muslims.

The current study also demonstrates that linguistic Bayesian phylogenetics (LBP) has no inherent bias to produce older dates for language dispersals; while the origin of the Indo-European (IE) language family has been dated to the early European Neolithic, and now Semitic to ~6,000 years, the spread of Melanesian languages to Pleistocene times, and of the Austronesian settlement of the Pacific to ~5,000 years. The congruence between LBP and traditional archaeology in all these cases should force IE exceptionalists who cling to the old theory of "steppe horse riders" to explain why, only in the dispersal of IE, it should LBP should have failed.

The paper also has free supplementary data, including a multistate phylogeny (pdf) of Semitic languages (reproduced top left of this post).

(More details to follow after I thoroughly read the paper)

UPDATE (Aug 27):

From the paper:
Furthermore, Eblaite (no Eblaite wordlists were available for our study), the closest relative of Akkadian and the only other member of East Semitic, was spoken in the Levant (specifically the northeast Levant or present-day Syria; Gordon 1997), which is also where some of the oldest West Semitic languages were spoken (Ugaritic, Aramaic and ancient Hebrew). The presence of ancient members of the two oldest Semitic groups (East andWest Semitic) in the same region of the Levant, combined with a possible long interval (100–3000 years) between the origin of Semitic and the appearance of Akkadian in Sumer, suggests a Semitic origin in the northeast Levant and a later movement of Akkadian eastward into Mesopotamia and Sumer (see figure 1 for a map of our proposed Semitic dispersals).
An origin of Semitic in northeast Levant (Syria) would be consistent with the observed east-west cline of decreasing J1 frequency in the Levant; the authors do, however, mention that the possibility for unknown extinct languages of the Semitic language may shift both the age of the language and its place of origin.
Lacking closely related non-Semitic languages to serve as out-groups in our phylogeny, we cannot estimate when or where the ancestor of all Semitic languages diverged from Afroasiatic. Furthermore, it is likely that some early Semitic languages became extinct and left no record of their existence. This is especially probable if early Semitic societies were pastoralist in nature (Blench 2006), as pastoralists are less likely to leave epigraphic and archaeological evidence of their languages.
A pastoralist association of Semitic languages is also consistent with the observed correlation of haplogroup J1 with herders and J2 with settled farmers in the Near East.


Proc. R. Soc. B 7 August 2009 vol. 276 no. 1668 2703-2710

Bayesian phylogenetic analysis of Semitic languages identifies an Early Bronze Age origin of Semitic in the Near East

Andrew Kitchen et al.

Abstract

The evolution of languages provides a unique opportunity to study human population history. The origin of Semitic and the nature of dispersals by Semitic-speaking populations are of great importance to our understanding of the ancient history of the Middle East and Horn of Africa. Semitic populations are associated with the oldest written languages and urban civilizations in the region, which gave rise to some of the world's first major religious and literary traditions. In this study, we employ Bayesian computational phylogenetic techniques recently developed in evolutionary biology to analyse Semitic lexical data by modelling language evolution and explicitly testing alternative hypotheses of Semitic history. We implement a relaxed linguistic clock to date language divergences and use epigraphic evidence for the sampling dates of extinct Semitic languages to calibrate the rate of language evolution. Our statistical tests of alternative Semitic histories support an initial divergence of Akkadian from ancestral Semitic over competing hypotheses (e.g. an African origin of Semitic). We estimate an Early Bronze Age origin for Semitic approximately 5750 years ago in the Levant, and further propose that contemporary Ethiosemitic languages of Africa reflect a single introduction of early Ethiosemitic from southern Arabia approximately 2800 years ago.

Link

204 comments:

1 – 200 of 204   Newer›   Newest»
Jean said...

It is rather for those few who accept the controversial Atkinson and Gray date for PIE to explain how a language of 7,800 and 9,800 years BP (their date for PIE) could contain words from a language that did not exist until 5,750 BP (the new date for Proto-Semitic.)

The most charitable thing to say of Atkinson and Gray's paper, which was ripped apart by linguists on publication, is that it was an early essay in a new method. This new study has evidently had the benefit of far more input from linguistics, and is much more plausible.

The success of a new method must be judged by how well the result fits expectations from other evidence.

eurologist said...

On the Anatolian origin of IE, it rests on the week contribution of Tocharian and Hittite. A Danubian origin (my favorite) 7,500 to 8,500 years ago actually fits their data pretty much as well, and is more plausible.

The Kurgan carriages got stuck in the muddy and densely forested European NE and came way too late, anyway.

Dienekes said...

It is rather for those few who accept the controversial Atkinson and Gray date for PIE to explain how a language of 7,800 and 9,800 years BP (their date for PIE) could contain words from a language that did not exist until 5,750 BP (the new date for Proto-Semitic.)

Semitic _existed_ in 5,750BP; its _breakup_ can be dated to that time, not its existence. Plus, an origin of PIE in Asia Minor explains the Semitic loan words a hell of a lot better than one north of the Black Sea.

which was ripped apart by linguists on publication

All the so-called criticisms have been addressed by Gray and Atkinson in subsequent publications.

Maju said...

I find quite unsurprising that Semitic is dated to c. 4000 BCE, as that is the time they appear in History. Also Semitic derives from a deeper linguistic family, Afroasiatic, that should not be much older than 8-10,000 BCE.

But I find surprising that this is used to reclaim the Anatolian hypothesis for Indoeuropean. Why were there so many non-IE languages in Anatolia and the other areas affected by the Neolithic (J2 if you wish) wave? From West to East: Tartessian, Iberian, Etruscan, Eteocretan, Eteocypriot, Hattic, Hurrio-Urartean, Caucasian languages (3 families), Sumerian, Elamite, Dravidian... some of them still alive and kicking.

Dienekes said...

Why were there so many non-IE languages in Anatolia and the other areas affected by the Neolithic

Why shouldn't there be? PIE was simply one of the languages of the Neolithic peoples, it was not -and was not claimed to be- the exclusive language of the Neolithic. Its good fortune was that it was the one that expanded into Europe where it outcompeted the languages of most of the pre-agricultural populations of the continent.

A very good analogy would be with the spread of Spanish into the New World. Spanish co-existed with many other languages in Europe, but it had the good fortune to expand into new territory occupied by technologically more primitive peoples, and thus achieved there (in the New World) the domninance that it could not have achieved in Europe.

Maju said...

Its good fortune was that it was the one that expanded into Europe where it outcompeted the languages of most of the pre-agricultural populations of the continent.

Outcompeted how?

And how do you explain the same process happening in South Asia or the Steppes?

If you told me that proto-proto-IE somehow reached the steppes (Samara valley probably) from an expanding multilingual Neolithic core, then I could look at it from another perspective. Just do not forget that all extant IE languages (excepting Albanian, Greek and Armenian) belong to just two subfamilies, probably liked originally at the time of Samara-Yamna. I could consider that languages like Albanian or Tocharian, that seem very old in the context of IE diversification (in all trees I've seen - Greek and Armenian vary instead) could have spread from a pre-Kurgan origin, maybe associated to secondary Neolithic waves.

But, even then, it stands highly suspicious that only IE languages survived the process, with so much diversity at the Neolithic spread. If all European IE languages would be Western IE, then we could look at some specific pattern. But all the other surviving languages (except Basque and Uralic) are also IE (but not Western IE).

We therefore need an explanation for the spread of IE (as such, not this or that subfamily) and a Neolithic with dozens of unrelated linguistic families doesn't explain anything at all.

Dienekes said...

Outcompeted how?

Demographically.

And how do you explain the same process happening in South Asia or the Steppes?

The steppes did not have substantial populations before the domestication of the horse, which allowed their exploitation.

As for South Asia, its IE-ization occurred ~1,500BC, when social complexity was already sufficient to allow for language spread via elite dominance.

But, even then, it stands highly suspicious that only IE languages survived the process, with so much diversity at the Neolithic spread.

Who told you that only IE languages survived? as recently as antiquity Europe was home to several non-IE languages; their subsequent demise has more to do with the Greco-Roman expansion, and perhaps the success of the Celts/Slavs/Germans further north, although that is harder to see due to the pre-literate nature of those societies. Basque and Finno-Ugrian may be the only ones left today, but that was not the case two thousand years ago.

Unknown said...

Genetic, Archaeological evidence points to east. Kurgan or Non Kurgan
Only for Languages Bayesian phylogeny provided some hope.
"PIE in Asia Minor explains the Semitic loan words a hell of a lot better than one north of the black sea"
Semitic language or people are just small minority living parallel to IE . Loan words are common either way. Semitic languages have far more IE words relatively please count. Known thing in Iran, Arabia, India.
North or south of the black sea does not matter if you are talking Neolithic peoples.
is there any proof of expansion to east from Asia Minor where IE languages also exist. So while people are migrating west somehow language travelled east. If you say demographic it has to spread both sides.
Semitic( desert area) contribution to the world in Pre bonze age for language, Culture, Industry , genetic is very negligible as populations are small and contact is less. Most of language, script, industrial development in non Semitic areas only
They are just living in desert between Afro asiatic and IE and language naturally has both the words.
You said Languages spread thru demography. That's why genetic markup and languages co exists parallel lot of times. Which is mostly true to IE? Same proof can be taken in your Basque and Finno-Ugrian example. Genetics are different.
Again Finno-Ugrian has far more similar words,deities in Indian lanaguages.
You are again wrong about south Asia. Still it is the place of highest population of Non IE languages in Eurasia. 300-400 millions of Dravidian language people with majority Y haplo R1a+R2+H. IE is restricted to north only. there is no demic diffusion or elite dominance . Just demographic spread along with language a later wave of R1a. Separate cultural or genetic people talk separate languages Semitic or IE or Dravidian or North Indian. We cannot argue demographic for one argument and use Bayesian phylogeny for another.

Andrew Lancaster said...

Maju, though I tend to be sceptical myself about the early Asian origins theories for PIE, you have to be careful how much you lean on IE being so European in recent millenia. I think we can be quite certain that Celtic and Latin and Germanic and Slavic spread from smaller homelands, that might have all been quite close to each other compared to the enormous distribution they had later. In other scenarios IE is seen as having spread from a smaller area in Eastern Europe, but then eventually from further east. That is not so different from the Anatolian theory, where they also spread from Eastern Europe, but previously from the south.

Maju said...

"Outcompeted how?"

Demographically
.

That's not a valid answer: they were supposed to be all Neolithic agriculturalist peoples.

You are not presenting any model. Renfrew's hypothesis at least says that IEs moved out of Anatolia and benefitted from their exclussive control of a new technology:; agriculture. But now you're saying they also outcompeted other agriculturalists by mere demography! Think again.

In fact, in all the cases mentioned that we know how the pre-IE agriculturalist were "outcompeted" by IEs (or in other cases by Semites), it was by the force of arms: they were dominated military and politically and then absorbed culturally.

Who told you that only IE languages survived? as recently as antiquity Europe was home to several non-IE languages...

Precisely! You had dozens of apparently unrelated linguistic families: what made specifically Indoeuropeans so successful if they were nothing but farmers like the rest.

They were already successful in Antiquity: in fact the other languages we know of were already remnants when they arrived to history and vanished soon after. And IEs even conquered and absorbed each other way too often (Semites too).

Would it have been a random process, each of the nations you mention would speak a distinct language, not all IE/Semitic (and later Turkic).

What do IEs, Semites and Turks have in common? What made them so succesful? AFAIK two related elements: (1) steppary/semidesert nomadism (they were virtually unconquerable there and could take whatever opportunity to vanquish their sedentary neighbours) and (2) ideology/religion of success (amoral/unethical).

Btw, what happened to Indo-Uralic in all this hypothesis of Anatolian origins? Indo-Uralic is the closest thing to a demonstrated linguistic superfamily for IE and Uralic but it almost necesarily implies a continental (steppe, tundra) origin or at least coalescence for PIE and Proto-Uralic.

Even if it's only a sprachbund, nobody suggests an alternative Indo-Hattic or Indo-Sumerian or Indo-Hurrian or Indo-Tyrrenian. Wonder why.

Maju said...

Maju, though I tend to be sceptical myself about the early Asian origins theories for PIE, you have to be careful how much you lean on IE being so European in recent millenia. I think we can be quite certain that Celtic and Latin and Germanic and Slavic spread from smaller homelands, that might have all been quite close to each other compared to the enormous distribution they had later. In other scenarios IE is seen as having spread from a smaller area in Eastern Europe, but then eventually from further east. That is not so different from the Anatolian theory, where they also spread from Eastern Europe, but previously from the south.

But when you analyze the archeology of Central Europe, you see an alien transformation in the Early/Middle Chalcolithic: Corded Ware (probably at the origins of Western IE) is not Danubian: it is Kurgan; and there was an easy to discern process by which it was estabilished in the previous milennium and by which Danubian populations were conquered and/or culturally absorbed into that Central European Kurgan cultural complex. Even when you first hear from Greek-speakers in Greece, you see kurgans (tumuli) too showing up in the archaeological record and leading directly to Mycenaean Greece a few centuries later.

The Anatolian hypothesis fails at the crucial element: if the supposed Anatolian IEs spread with Neolithic and benefitted of agriculture as their main advantage (and not their bellicosity), why didn't the other Neolithic linguistic families have the same success? Why did they lose, not only in Western Europe but in the Balcans, Anatolia, Eastern Europe, Iran and even such a remote place as South Asia?

I would expect regionalized patterns of success, not this continental homogeneity. Or, if IEs were the only Neolithics, I would expect no other languages showing up.

Furthermore there are serious linguistic considerations: agricultural terms are generally not pan-IE, while those related to the steppes seem to be. Plus all other chronological estimates I know of suggest a timeline much more recent than Neolithic: c. 6000 years ago.

Andrew Lancaster said...

Maju, you ask Dienekes why one Neolithic Asian language would do better in a New World than another. Dienekes gave a pretty good example of Spain being the one language to conquer the new world. So it is possible. You ask for a model, but what model can you have? Maybe there were a few ambitious leaders?

I think you are on stronger ground when you say "precisely!" to Dienekes pointing out that IE may have been a minority in classical antiquity. What it implies is that we have historical reasons for thinking IE did a lot of its spreading not long before classical antiquity comes into focus in the written record. If this is the case then why associate it with farming? The farming revolution came and went and the modern IE families were all apparently fairly localized long after it.

In other words, working backwards from what we know in historical records has been a big part of why people argue for the IE family being younger than the Neolithic.

I think more sophisticated Anatolian origins theories normally propose that there was only a first small initial wave in early times, perhaps not getting far into Europe at all, but perhaps at first just into part of the Balkans and the Ukraine. You still then need a motor to power the later expansions such as the big Indo-Aryan one, and the one which apparently pushed a branch into Western Europe.

Regards
Andrew

Dienekes said...

But now you're saying they also outcompeted other agriculturalists by mere demography! Think again.

What other agriculturalists? In the Renfrew model, the Neolithic of Europe was Indo-European speaking. That is not incompatible with the arrival of other non-IE languages from the Near East or the survival of non-IE languages from Mesolithic Europe.

Europe became IE, because its major Neolithic source population (from Central Anatolia) was IE. This is why the Americas became English/Spanish/Brazilian speaking, because they happened to be colonized by those areas of Europe.

It is also quite reasonable that -even if many language families participated in the Neolithic colonization of Europe- that one of them would have outcompeted the others by a great margin; Spanish has outcompeted French in the New World for example.

You had dozens of apparently unrelated linguistic families: what made specifically Indoeuropeans so successful if they were nothing but farmers like the rest.

Without discounting the importance of luck, the major factor seems to have been opportunity; the Anatolian Indo-Europeans happened to be placed in an ideal location to spread their languages to Europe.

Was there any "technological" reason why Spaniards and Englishmen colonized the New World, and not the multiple other languages of Europe? Not really, they just happened to border the Atlantic.

agricultural terms are generally not pan-IE

To quote Renfrew, here is no IE term for hands, does that mean PIEs didn't have hands? The non-existence of PIE terms for agricultural terms could simply be a random event (as the disappearance of the term for hands), aided, perhaps, by the fact that agricultural products (unlike e.g., sheep or cattle) are very tied to their ecological zone. How can you expect to have terms for Mediterranean plants in Tocharian or Indo-Aryan?

Kepler said...

Dienekes,
Do you know of any reseach on old DNA in places such as Palestine or Northern Africa? I suppose there is little. As I wrote in another post, an archaeologist told me there is a project that will start this year where there will be some DNA analysis of very old material in Palestine/Israel...but we will have to wait.

Iberian Peninsula's languages:
http://en.wikipedia.org/wiki/Paleohispanic_languages

Spain invasion in Latin was very radical. You can see the dramatic change that took place in may ways.

Obviously, technological differences and epidemics were very strong there, in Europe there was not such a thing. Still: would we not expect some stronger traces of IE expansion?

In the Americas the predominance of European Y-haplogroups is overwhelming. Is it possible to see if something like that happened in Europe?

Maju said...

Maju, you ask Dienekes why one Neolithic Asian language would do better in a New World than another. Dienekes gave a pretty good example of Spain being the one language to conquer the new world. So it is possible.

But that is perfectly explained because the Castilians were in America almost 100 years before anyone else and kept hold of most of it for many centuries later. It's not easy to compare anyhow. I think in comparisons like Austronesians, Turkics or Semites: there you can build up a model of expansion that is consistent.

You ask for a model, but what model can you have? Maybe there were a few ambitious leaders? -

That's precisely the wrong answer because that implies some sort of radical cultural differences: different origin therefore.

Also, it is no model: you claim that a culture/language/identity expanded through half the largest continent only by chance. Sorry but I do not believe in coincidences of such huge dimensions.

In other words, working backwards from what we know in historical records has been a big part of why people argue for the IE family being younger than the Neolithic.

Sure. That's one of the most clear evidences against Renfrew's model: that nobody seems to have spoken IE in Anatolia before the Hittites arrived (or elsewhere mutatis mutandi).

I think more sophisticated Anatolian origins theories normally propose that there was only a first small initial wave in early times, perhaps not getting far into Europe at all, but perhaps at first just into part of the Balkans and the Ukraine.

Makes no sense. One would expect all Balcanic Neolithic peoples (and by derivation possibly all Danubian Neolithic and Cardium Pottery Neolithic ones) to speak the same type of language. They are way too close culturally to be so different.

Also the Ukraine that matters for the Kurgan model (I understand you're trying to reach there) held an unrelated Neolithic culture (Dniepr-Don) that arguably is derived from the local Epigravettian and in any case it is unrelated to the Balcanic or Danubian Neolitic. Only SW Ukraine was in the Balcanic/Danubian sphere, together with the two Moldavias.

Ardagastus said...

Gray & Atkinson's papers are here and here's Dyen, Kruskal and Black lexical database (1997).

And some of my reasons to doubt are:

- G & A propose a number of assumptions which are not granted: that cognate sets are enough to measure linguistic evolution ("cognate sets constitute discrete, relatively unambiguous heritable evolutionary units with a birth and death and there is no reason to suppose they are any more or less fundamental to language evolution than semantic categories", 2006:94), that linguistic change is reasonably similar in time and space ("our model assumes that the appearance and disappearance of cognates is randomly distributed about some mean value", 2006:95), and even when they attempt to refute criticism they dodge the burden of proof which is arguably theirs. To support their conclusions they must prove their argumentation is not only valid, but also sound. For more about these assumptions, read also Bill Poser's Dating Indo-European

- the lexical database of Isidore Dyen et al. is not bulletproof. For instance, the Romanian ANIMAL and the English WING are listed as cognates, though they are later loanwords. On the other hand, for BELLY this database gives Romanian PINTEC(E) (from Latin PANTEX,-ICIS), but not VINTRE (from Latin VENTER,-RIS) which would count as cognate with other Romance words. G & A's modeling does not deal with these errors and they consider "Dyen et al. (1992) Indo-European lexical data base [...] contains expert cognacy judgements". However when checking the bibliographical support we will find even materials such as pocket dictionaries and conversational guides, while many reliable etymological dictionaries and tools (necessary for "expert cognacy judgements") are missing.

- the case of Latin and Romance shows also an interesting aspects of lexical change. There are some Latin words with similar meanings which were inherited in a different way in various Romance languages, thus the dichotomy cognate vs non-cognate (further interpreted in terms of change) might be misleading in assessing the chronology or even in building a tree-model. That is the case with Latin PANTEX and VENTER, but there are many other examples

- I'm using Romance linguistic group also to verify their model against historical and linguistic data. The Romance sub-tree (2006:99, figure 8.8, the first large family) looks rather unpersuasive. Some splits like between Romanian and Vlach languages or between Spanish and Portuguese happened much earlier than predicted by their model, and many Romanists would doubt the existence of a Romanian-Vlach-Ladin, Galo-Iberian, or even Romance-but-Italian-and-Sardinian languages in the second half of first millenium AD.

Maju said...

What other agriculturalists? In the Renfrew model, the Neolithic of Europe was Indo-European speaking. That is not incompatible with the arrival of other non-IE languages from the Near East or the survival of non-IE languages from Mesolithic Europe.

Europe became IE, because its major Neolithic source population (from Central Anatolia) was IE
.

AFAIK Central Anatolia was Hattic-speaking (pre-IE, possibly related to NW Caucasian).

Also I have never ever read that Balcanic Neolithic is derived from Central Anatolia but actually from Southern Anatolia: Hacilar specially but even Catal Hoyuk is mentioned at times. Of course, this may be biased because of patchy archaeological research in Turkey but claiming otherwise demands evidence.

Sorry that I just realized that Renfrew doesn't even posit a parallel eastward march of IEs towards India, maybe associated to the J2b found there in high amounts and that could be of Neolithic origin. It seems he fancied that Indo-Iranian appeared in the Balcans (why?) and then the farmers conquered the steppes somehow before leading to South Asia. A total nonsense!

It is also quite reasonable that -even if many language families participated in the Neolithic colonization of Europe- that one of them would have outcompeted the others by a great margin; Spanish has outcompeted French in the New World for example

Erm. English has outcompeted French (in Canada, Great Lakes, Louisiana), Spanish was there before either one arrived. Spanish is the most spoken language of America because it's also the oldest of colonial languages by a significative margin. America is anyhow a bad comparison: too modern, rapid and overwhelming. And anyhow all these languages are Western Indoeuropean, none is, say, Semitic or Turkic or even Uralic.

Without discounting the importance of luck...

Loads of luck. Too much to be credible.

...the Anatolian Indo-Europeans happened to be placed in an ideal location to spread their languages to Europe.

Sure. The Anatolians whatever surely spread their languages to Europe, with founder effect in Northern Greece probably, just like E1b1b1. But how do you explain their arrival to India, Uyghuristan, etc.? How do you explain that Anatolians were speaking other languages, not Indoeuropean and that was also the case with other possibly (in many cases surely) Anatolian-derived languages like Etruscan, Eteocretan, etc.

How do you explain the lack of pan-IE cognates for Mediterranean vegetation or agricultural terms?

Was there any "technological" reason why Spaniards and Englishmen colonized the New World, and not the multiple other languages of Europe? Not really, they just happened to border the Atlantic.

Hmmm... there were power reasons: the Habsburgs were the largest power of Western and Central Europe in the time they colonized America and they were replaced by England later on. This is more a military/economic history of imperial states rather than that of tribesmen.

Whatever the case: you have a clear straight archaeological pattern there too. There is no cultural replacement as happened with Danubians in Central Europe. Danubians were just supressed as a distinct culture c. 2400 BCE. Most of that suppression was caused by the Kurgan peoples (the real IEs), though Aquitanians also did their share from the SW, thanks to the massive use of some kind of very effective bow weaponry.

Maju said...

(cont.)

To quote Renfrew, here is no IE term for hands, does that mean PIEs didn't have hands? The non-existence of PIE terms for agricultural terms could simply be a random event (as the disappearance of the term for hands), aided, perhaps, by the fact that agricultural products (unlike e.g., sheep or cattle) are very tied to their ecological zone. How can you expect to have terms for Mediterranean plants in Tocharian or Indo-Aryan?-

Did not IVC plant wheat or other cereals?

But never mind, since I found, just now, that Renfrew claims a European, specifically Balcanic, origin for Indo-Iranian, I have lost whatever respect I had for that model.

Simply put: it'd be easier to claim that IEs migrated from the Highland Fertile Crescent to South Asia directly, with all that J2b, via Iran. That could make some (not too much but certainly much better) sense.

But claiming that IEs migrated from the Balcans to the Steppes is just in total contradiction with archaeology and genetics. And only shows how necessary is the Kurgan model even for those that reject it.

pconroy said...

But of course we do have a mechanism other than - as Maju suggested - warlike tendancies/agression and organized religion, there is also:

1. Iron weapons

2. LCT - Lactase Persistence gene - helping carriers get something on the order of 10-20% more nutrition from the same quantity of dairy products than non-carriers.

This could mean that a small IE warband with superior weapons, martial skills and cohesion could overpower a larger population. Then their Y-DNA would spread with each subsequent generation in a cumulative fashion.

The coalescence time for R1b in Western Europe is on the order of 4,000 years - which is about right.

As regards the homeland of the PIE speakers, we can infer that it bordered on the Uralic and the Semitic language groups - therefore somewhere from Eastern Anatolia to the North Pontic Steppes.

Maju said...

Do you know of any reseach on old DNA in places such as Palestine or Northern Africa? -

Apart of the recently published here aDNA studies of Guanches, there is Kefi'05 on Taforalt ancient mtDNA (quite similar to modern Riffian, btw).

Maju said...

This could mean that a small IE warband with superior weapons, martial skills and cohesion could overpower a larger population. Then their Y-DNA would spread with each subsequent generation in a cumulative fashion.

But you are talking of the warlike bands of steppe riders, right? Otherwise I make no sense.

Anyhow Danubians, we like it or not, were smashed. They did not expand: they contracted and vanished. And Kurgan cultures replaced them essentially.

LCT - Lactase Persistence gene

Which is much higher among Basques than among their IE neighbours. Please!

The coalescence time for R1b in Western Europe is on the order of 4,000 years - which is about right.

Even if that would be correct, that is not, it would not be right: IEs only arrived to Britain c. 2300 years ago and Basques were never IEs. And anyhow, we should expect a cline from Central Europe to the West and SW not the other way around. Something else overlaid Central European R1b and replaced it with other stuff. These were at least in part IEs, as attested by a decrease of R1b in the areas more intensely and repeatedly Indoeuropeized of Western Europe too. R1b decreases towards the origins of IEs, it just cannot be IE.

As regards the homeland of the PIE speakers, we can infer that it bordered on the Uralic and the Semitic language groups - therefore somewhere from Eastern Anatolia to the North Pontic Steppes.

I am not convinced by any alleged relationship between Semitic/Afroasiatic and IE: it's mere shared borrowing of pre-existent Neolithic terms of some other language family.

Maju said...

Gray & Atkinson's papers are here and here's Dyen, Kruskal and Black lexical database (1997).

Thanks, Ardagastus. I also share your criticisms. How can be Romanian closer to Spanish than Italian, when Italian is perfectly understandable for a Spanish-speaker and vice versa, while Romanian sounds like some sort of unintelligible "Russian"? Also all standard classifications would say that there is a Gallo-Romanic that includes Italian and French, nowhere to be seen in this scheme.

Similarly there is consensus that Frisian and English are cousins but this tree places Frisian closer to Dutch.

And well, I did not know that Greek is only 800 years old. Ahem!

Gioiello said...

Sometime acts the theory of the “aree laterali” of Bartoli, like the Indo-European for “king”: Latin “rex”, Gallic “rix”, Sanskrit “raja”; so we have Spanish “hermosa” and Rumanian “frumoasa” from Latin ”formosa”, what in Italian we say “bella”, but we have also “formosa”, but in another meaning.
Anyway Spanish is closer to Italian than Rumanian, even because we Italians familiarize with Spanish by children, what doesn’t happen with Rumanian.

Ardagastus said...

And well, I did not know that Greek is only 800 years old. Ahem!

Yes, the Greek branch looks strange in G & A's tree-model. I assume those Greek idioms are actually representing the word sets selected from some dictionaries, but the chronology is unfactual and that is obvious for Katharevousa and Dimotiki (there was no such split 800 years ago)

Kepler said...

First of all: people talk about Sprachbund usually between unrelated or distantly related languages coming closer, but what about closer ones? (or is there a term for that case, closer languages getting closer during special periods?, I can't recall)

Romanian has some weird things like articles at the end - domul for man-the (shared with Bulgarian) and lots of Slavic words, like vremea, da,etc. Obvious it is still a Romance language and Romanians learn Italian and Spanish rather fast, but Italian is closer to Spanish. The lexical borrowing between Spanish and Italian has been continuous, whereas Romanians were isolated for a huge part of their history.

I think people have never developed consistent ways of measuring closeness between languages.
One of the fetishes some linguists have is with how languages create plural and details like that (Romanian mostly with -i, -e, more similar to Italian, which took it from Nominative Latin, while we took it mostly from accusative forms of the plural)

With regards to Frisian: I speak some Dutch and even though it is hard, I can understand written Frisian mostly thanks to Dutch, I don't think thanks to English (which has so many Romance words)

Frisian (Wikipedia):
"Yn 'e lêste helte fan 'e njoggentjinde en de earste helte fan 'e tweintichste ieu is de Fryske standerttaal gearstald."
Dutch something like:
"In de latste helft van de
negentiende en de eerste helft
van de twintigste eeuw is het Friese standaardtaal ontstaan" (I would say ontstaan, gearstald sounds like participle of herstellen, produce, not sure)

Dienekes said...

But that is perfectly explained because the Castilians were in America almost 100 years before anyone else and kept hold of most of it for many centuries later. It's not easy to compare anyhow.

And Indo-European was probably the first farmer language in Europe; by the time anyone else cared to head west, it was already blooming demographically.

though they are later loanwords.

In "Does horizontal transfer invalidate cultural phylogenies" Greenhill, Curie, and Gray show that horizontal transfer between languages tends to underestimate divergence times, although the method is robust to realistic levels of it.

Some splits like between Romanian and Vlach languages or between Spanish and Portuguese happened much earlier than predicted by their model

But this goes against the criticism that they overestiate linguistic time depth.

And well, I did not know that Greek is only 800 years old. Ahem!

Apparently you are unable to understand a simple chart. Greek diverged from other IE languages a long time ago; 800 years is the inferred time of separation of the included dialects, which is understandable as Greek is a very conservative language which has undergone fewer changes in the historical period than other languages.

1. Iron weapons

No one seriously considers an iron age spread of Indo-European

LCT - Lactase Persistence gene - helping carriers get something on the order of 10-20% more nutrition from the same quantity of dairy products than non-carriers.

And farming allows a 100-fold population density than hunting and gathering, a much greater demographic advantage. Not to mention that LCT carriers are able to digest milk but non-carriers have no trouble getting all the nutrition from secondary products of milk which can be preserved for a longer period of time in a hotter climate.

Dienekes said...

Some of my Vlach acquaintances tell me that they understand Italian easily. In any case mutual comprehensibility is as much a function of phonology as it is that of number of shared words.

Dienekes said...

that is obvious for Katharevousa and Dimotiki (there was no such split 800 years ago)

An 800 years inferred difference between Dimotiki and Katharevousa sounds perfect, since Dimotiki is the natural evolution of the language, and Katharevousa is about halfway between koine and Dimotiki, which are separated by about two millennia.

Kepler said...

Guys,

Spaniards really spread their genes in the Americas in a huge way.
Even some of the places with the highest amount of Indians, Peru, you have something like this:

http://content.karger.com/ProdukteDB/produkte.asp?doi=10.1159/000022964

So: even in a Quechua group in Peru half the males had haplogroups of European origin. In Venezuela and Argentina and other places on the male side ratios go to 90-92%.
mtDNA from Europe is less represented, but still: it is about 9% in this group, it goes to around 40-50% in Argentina.

Was there something of the sort in Europe? I think not.

Ardagastus said...

But this goes against the criticism that they overestimate linguistic time depth. Actually my objection is that their method is flawed, that cognate sets are not a way to estimate the chronology of language splits (see my PANTEX/VENTER example), that Dyen database contains both false cognates and unrecognized cognates and moreover that their tree is arguably unfactual when verified against recent linguistc history.

An 800 years inferred difference between Dimotiki and Katharevousa sounds perfect, since Dimotiki is the natural evolution of the language, and Katharevousa is about halfway between koine and Dimotiki, which are separated by about two millennia. Not at all, Katharevousa was created in the late 18th century and this tree fails to predict that. In many cases we will find a dual system of literary (Katharevousa, Attic Greek, Classical Latin, many modern standard linguistic varieties) and vulgar language, one of the many linguistic realities which this model doesn't take in account. Even if one would fall for that "halfway" definition, Koine Greek was created in the Hellenistic age, more than 2 millenia ago.
And this is not so much about the inabilities of Maju but of Gray, Atkinson, Dyen and their colleagues. Because the five Greek "dialects" represented in this tree are actually the word sets given by those five modern Greek dictionaries (see my previous comment). And while Greek ML (Modern Lesbian Greek) might be considered a separate dialect of Greek, I don't think Greek Mod (Modern Greek) or Greek MD (Modern Spoken Greek) are. The latter two are either Katharevousa or Dimotiki. Gray and Atkinson's reconstruction is ridiculous, they seem to believe the data from five Greek dictionaries actually represent five dialects which evolved in time from a common root. They predicted a chronology where there was none.

Dienekes said...

Not at all, Katharevousa was created in the late 18th century and this tree fails to predict that.

How do you expect any method to predict when an artificial language was created? Would you expect a method to predict that modern Hebrew was created in the 19th century?

However, katharevous as an artificial language is intermediate between koine and demotic Greek, and a split time of 800 years makes perfect sense.

Anyway, all the differences between the various Greek dialects are small change compared to the long branch that links Greek to the rest of the IE family, so nothing they could have done in that part of the tree would significantly affect their conclusions.

Ardagastus said...

How do you expect any method to predict when an artificial language was created? Would you expect a method to predict that modern Hebrew was created in the 19th century?

However, katharevousa as an artificial language is intermediate between koine and demotic Greek, and a split time of 800 years makes perfect sense.

Historical linguists already know when Katharevousa was created, so I guess there are some methods to detect massive changes in the history of some language, even when it's about borrowings from an ancient variety of it. The Greek Koine was also more or less artificial (as you put it), though a different phenomenon. I believe all standard written languages are more or less artificial, and this is one of those several socio-cultural factors influencing the lexical replacement rate, possibly in unexpected ways. Moreover, from a point of view relying exclusively on words, there's no essential difference between modern standard Romance languages and modern standard Greek, as the former borrowed consistently from Medieval and Ecclesiastical Latin.
But G & A claim they have reliable sets of cognates, which should filter out all what's not naturally inherited. They also provide no additional scaling tools, so 800 years should be read as 800 years.

Anyway, all the differences between the various Greek dialects are small change compared to the long branch that links Greek to the rest of the IE family, so nothing they could have done in that part of the tree would significantly affect their conclusions.

When we found out that Earth is not flat, few if any thought "well, perhaps some other planets are flat". If their method fails to achieve good results for the history we already know, how can we trust it for a prehistory we don't know? How can they prove their conclusions?

Dienekes said...

When we found out that Earth is not flat, few if any thought "well, perhaps some other planets are flat". If their method fails to achieve good results for the history we already know, how can we trust it for a prehistory we don't know? How can they prove their conclusions?

I think you are nitpicking. An inclusion of an artificial language in the analysis would indeed bias the results, but this would only happen if they included a new basal branch in the tree that was more different from the rest of IE than Hittite is from the rest. The inclusion of Greek dialects has no effect on the age estimate, as these are all very closely related to all the rest.

Written languages are perhaps always slightly artificial, but we only have written languages to work with.

Maju said...

Greek diverged from other IE languages a long time ago; 800 years is the inferred time of separation of the included dialects, which is understandable as Greek is a very conservative language which has undergone fewer changes in the historical period than other languages.

So languages can evolve at different rates? And that may cause any deductive chronology to fail catastrophically... hmmm, interesting to see that idea I had somewhat confirmed.

Anyhow, I would expect such a cosmopolitan language as Classical Greek to have evolved very fast in Antiquity (lots of influences and new speakers introducing neologisms and idioms), even if later the rate of evolution slowed somewhat.

But this goes against the criticism that they overestiate linguistic time depth.

It shows clear inconsistencies in the method or at least the results: it falsifies the hypothesis and therefore causes it to be invalid scientifically.

And Indo-European was probably the first farmer language in Europe; by the time anyone else cared to head west, it was already blooming demographically.

Very conjectural. Anyhow, it would not explain expansion to the Steppes and Southern Asia. Danubian culture only existed between Northern France and SW Ukraine. And, as I said before, collapsed under mostly Kurgan pressure.

The so-called Anatolian model not just fails to account for the would-be extreme Danubization of Kurgan peoples in an illiterate context but, most crucially, fails to account for any way of expansion eastward that can be tested against archaeological data. Neither Danubians, nor Balcanic Neolithic peoples, nor Corded Ware ones ever expanded to the East, at least not beyond middle Russia (Corded Ware only). If you claim an steppe origin for Indoiranians and Tocharians, you need to cling to Samara-Yamna: Central European/Balcanic cultural processes can't account for that.

With regards to Frisian: I speak some Dutch and even though it is hard, I can understand written Frisian mostly thanks to Dutch, I don't think thanks to English (which has so many Romance words).

Yah, guess the huge amount of Romance words in English (and the sprachbund between Dutch/German and Frisian) may confuse the matter but serious linguists all classify Frisian closer to English than to anything else (based obviously on the Germanic words and structure and in what is known of pre-Norman Old English). Native English speakers often claim to be able to understand Frisian more or less after they overcome the pronunciation barrier. They never say that of Dutch or German.

Dienekes said...

It shows clear inconsistencies in the method or at least the results: it falsifies the hypothesis and therefore causes it to be invalid scientifically.


On the contrary, Gray has shown that their method is resilient to lexical borrowing and have studied how it functions in the presence of such borrowing.

If electrical conductivity increases in the presence of atmospheric moisture, we don't say that this renders the theory of electromagnetism invalid.

If you claim an steppe origin for Indoiranians and Tocharians, you need to cling to Samara-Yamna: Central European/Balcanic cultural processes can't account for that.

You seem to be quite selective in choosing your "missing links". You won't find any kurgans in South Asia (or Western Europe for that matter), yet you see no trouble in postulating ultimate Kurgan links to those IE languages.

In any case, we do not have to rely on either physical anthropology or archaeology to provide us with a chain of transmission going back all the way to Neolithic Anatolia; the presence of J2a in upper caste Indians is precisely such a clear and unambiguous link. Its correlation with caste status, its presence throughout India and the virtual lack of many other Middle Eastern Y chromosomes in that population make it all but certain that J2a is not a signal of occasional mixing in some Brahmin populations, but a real component in the ancestral Indo-Aryan gene pool.

Ardagastus said...

I think you are nitpicking. An inclusion of an artificial language in the analysis would indeed bias the results, but this would only happen if they included a new basal branch in the tree that was more different from the rest of IE than Hittite is from the rest. The inclusion of Greek dialects has no effect on the age estimate, as these are all very closely related to all the rest.


I already pointed out that according to their papers, they coded that tree only on what they considered naturally evolved characteristics (words), and they claim no bias at all in this regard. The lexical borrowings from older Greek idioms into modern ones are just borrowings (2006:93 : "Known borrowings were
not coded as cognate in the Dyen et al. data base" and there are several similar assertions that their data contains a small amount of borrowings)

However it's not only the Greek sub-tree which they have it wrong, it's also the Romance, possibly the Germanic (as pointed out in other comments), Albanian and probably some others. Get historical linguists to assess their work, not biologists, and you'll see the results.

Going back to prehistory, here's another refutation of G & A's chronology.

Dienekes said...

Once again, the presence of borrowings would make two languages to appear more similar to each other than they really are, and hence would underestimate their time of divergence.

Going back to prehistory, here's another refutation of G & A's chronology.

Nothing really new in this "refutation".

The only real argument I see in the linguistic evidence is the presence of "wagon terminology" in Proto-Indo-European. Let's discount -for the sake of argument- the possibility that the parts were known, but not the whole (after all, the parts of a clock had names before there was actually any clock). It still is a fact that wagons were invented early on in IE history, a few thousand years after the Neolithic colonization of Europe. At that time, the IEs were still geographically close to each other, and their differentiation had barely began -except in the case of Anatolian. So, the wagon terminology could easily spread among them.

Ardagastus said...

Once again, the presence of borrowings would make two languages to appear more similar to each other than they really are, and hence would underestimate their time of divergence.


But in the case of Greek they overestimated the divergence. The actual problem is that lexical change rates (and the chronology of linguistic evolutions derived from these) seem not to be those assumed by G & A.

Nothing really new in this "refutation".

It is new (2007) and it addresses also some arguments of G & A from their later papers (2006). The argument put forward by David Anthony is that wagon/wheel vocabulary cannot be borrowed after the split of PIE, because it's unlikely that after some 3000 years of linguistic evolution, proto-Greek, proto-Balto-Slavic, proto-Indo-Iranian, etc. had the same phonology as PIE. If the wagon/wheel vocabulary was a late borrowing sweeping through all IE languages, as you (and Gray) suggest, then we should see some linguistic evidence for such borrowings, and apparently there's none.

Also there's Maju earlier objection, which is also supported by Anthony, that Proto-Uralic was probably one of the neighbours of PIE.

Maju said...

If electrical conductivity increases in the presence of atmospheric moisture, we don't say that this renders the theory of electromagnetism invalid.

It would have us to change the circuitry so it can work independent of moisture. This is not the theory of electromagnetism but just a circuit that doesn't seem to work properly. Replace it.

You seem to be quite selective in choosing your "missing links". You won't find any kurgans in South Asia (or Western Europe for that matter), yet you see no trouble in postulating ultimate Kurgan links to those IE languages.

You seem to be ignoring the cultural aspects other than burial practices that were gradually altered within the expansion and just time passing within the Kurgan cultural complex. But, using archaeology, you can perfectly trace the cultural continuity anyhow.

Instead you and Renfrew lack of even a single archaeologically consistent explanation for the eastern part of your model.

In any case, we do not have to rely on either physical anthropology or archaeology to provide us with a chain of transmission going back all the way to Neolithic Anatolia...

I'd be happy enough if you could provide some more or less consistent conjecture on how the Balcano-Danubian Neolithic complex expanded to the East.

...the presence of J2a in upper caste Indians is precisely such a clear and unambiguous link.

J2a or J2b? The latter is the common one in both India and Europe AFAIK. J2a is rare outside West Asia.

Anyhow, I'd say (and many others too) that J2 in India looks Neolithic (i.e. direct migration from West Asia via Iran, also affecting Central Asia) long before the Indo-Iranian migrations.

Dienekes said...

because it's unlikely that after some 3000 years of linguistic evolution, proto-Greek, proto-Balto-Slavic, proto-Indo-Iranian, etc. had the same phonology as PIE.

There were none of these "proto" languages yet. The linguistic ancestors of all these later peoples were still all concentrated in a small part of Europe at that time.

J2a is rare outside West Asia.

Your facts are wrong.

Anyhow, I'd say (and many others too) that J2 in India looks Neolithic

Right, J2a is a Neolithic haplogroup that ended up concentrated in a particular social caste. I'd love to hear how a haplogroup that was around for a few thousand years in the population of India ended up at a different frequency in high vs. low castes.

Dienekes said...

Instead you and Renfrew lack of even a single archaeologically consistent explanation for the eastern part of your model.

No problem, as one of Renfrew's two hypotheses is identical to the Kurgan model for the eastern part of the IE expansion.

Ardagastus said...

There were none of these "proto" languages yet. The linguistic ancestors of all these later peoples were still all concentrated in a small part of Europe at that time.


This is not what the Gray & Atkinson tree shows (2006:99, figure 8.8). They claim a branch later becoming Hittite first split off (8700 years BP), followed by Tocharian (7900 BP), Graeco-Armenian (7300 BP), Albano-Indo-Iranian (6900 BP), Balto-Slavic (6500 BP), Albanian and Indo-Iranian (6500 BP), Greek and Armenian (6400 BP), Celtic (6100 BP).

Even if we disregard their tree and we just want to assess the Anatolian hypothesis, it's highly unlikely the language in 9000 BP sounded more or less the same as the same language in 6000 BP when the wheel/wagon vocabulary probably spread. Check also this post by Don Ringe.

Maju said...

No problem, as one of Renfrew's two hypotheses is identical to the Kurgan model for the eastern part of the IE expansion.

Yes, there is a problem: because Kurgan is oldest in Samara (by some 2000 years) than anywhere in Europe and there is no explanation of any sort given for how IE culture propagated from Central Europe or the Balcans to there.

That is a most central problem: that, like Hindutva explanations fails totally to look at the other half of the issue. It's just the new European (and Eurocentric) version of the Indian origins hypothesis, just change IVC for Balcano-Danubian Neolithic and is the same. Same central problem too: that the real, traceable, steppary IEs, who unavoidably connect South Asia and Europe, don't seem to have any connection whatsoever with either IVC or the BD Neolithic, much less an origin in any of them.

Dienekes said...

it's highly unlikely the language in 9000 BP sounded more or less the same as the same language in 6000 BP

First of all, Anatolian does not certainly have a wagon vocabulary. This is easily explained by the Renfrew hypothesis, as it stayed behind while the rest of the tree evolved out of Europe where the terminology evolved.

So, even if we were sure both about the position of Tocharian in the tree and its phonology, we just have an interval of 1,900 years during which it is quite plausible that the phonetic values for the wagon-related terminology were preserved. Hell, in Greek the phonetic value for axon was preserved for 3,000 years.

I would rather take G&A's formal quantitative approach based on cognates to subjective "expert" judgments on how quickly phonologies change.

Dienekes said...

Kurgan is oldest in Samara ... no explanation of any sort given for how IE culture propagated from Central Europe or the Balcans to there.

Well, Kurgans are completely absent from most of Europe (or Iran-India for that matter), so the steppe enthusiasts have a much bigger problem to tackle.

But, even the steppes provide an answer to the so-called problem of how Central European influence may have mobilized the earlier Kurgan populations.

Just as a handful of Altaic speakers ultimately linguistically converted and mobilized masses of Iranian speakers in Central Asia and sent them westward to Europe and Asia where they arrived as various types of "Turks", so it is entirely possible that a handful of IE speakers from the Balkan-Danubian-Central European area mobilized the Kurgan populations and sent them eastward to Asia.

To this, one might say "but what about the kurgans in Romania, Bulgaria, and Hungary" etc. that show an intrusive Kurgan population in the Balkans? But, this is not a problem at all; there are Scytho-Sarmatian kurgans all the way to Mongolia, but we all know whose language ultimately won out and spread over the steppe region.

Andrew Lancaster said...

Maju writes:

That's precisely the wrong answer because that implies some sort of radical cultural differences: different origin therefore.

Also, it is no model: you claim that a culture/language/identity expanded through half the largest continent only by chance. Sorry but I do not believe in coincidences of such huge dimensions.

I guess my point about having a good leader is that it could just be luck. And what I mean by that is that we do not need a model because there is no surprising coincidence needing to be explained. What is the coincidence you think needs to be explained?

Andrew Lancaster said...

Maju

You make some points which lead me to think that an important part of what you are arguing for is that the Balkan first farmers belonged to one culture, and that this culture likely came from further south than where we later find Anatolian languages.

I agree that the Greek and Cretan Neolithic show signs of possibly even coming by boat from the Levant. But the Cardial culture I think the origins are still a mystery, and for the Balkan Neolithic, I think the growing feeling is that it was quite separate from the Greek Neolithic, and may have come quite early from Anatolia.

Ardagastus said...

First of all, Anatolian does not certainly have a wagon vocabulary. This is easily explained by the Renfrew hypothesis, as it stayed behind while the rest of the tree evolved out of Europe where the terminology evolved.
That's easily explained in traditional PIE framework, as Anatolian branch was first to split off.


So, even if we were sure both about the position of Tocharian in the tree and its phonology, we just have an interval of 1,900 years during which it is quite plausible that the phonetic values for the wagon-related terminology were preserved. Hell, in Greek the phonetic value for axon was preserved for 3,000 years.

But around 6000 BP, according to that tree, we don't have only Tocharian, there are a number of different languages (see my previous post) and both explanations proposed by Gray and Atkinson (2006:102-3) were refuted by Anthony and Ringe. It is virtually impossible for the aforementioned vocabulary and phonology to remain so unitary in so many different languages after such a long time.

I would rather take G&A's formal quantitative approach based on cognates to subjective "expert" judgments on how quickly phonologies change.
Gray and Atkinson have no expertise in historical linguistics and they provided no proof whatsoever. Fomenko's new chronology is crap, formal approach is useless if the assumptions are wrong (and actually G & A rely heavily on "subjective expert judgements" in their assumptions)

Moreover, Ringe et al. used similar methods but starting from different assumptions and obtaining different trees.
Check, for example, this paper on methods of phylogenetic reconstruction or another one on IE trees (please note the chronology starting 6000 years BP!)

Dienekes said...

That's easily explained in traditional PIE framework, as Anatolian branch was first to split off.

A split involving movement of farmers from Anatolia to Greece (which is simple, and we know that it happened) is a hell of a lot better explanation than the wholly conjectural "first split" of Indo-Europeans all the way around the Black Sea to arrive in Anatolia.

But around 6000 BP, according to that tree, we don't have only Tocharian, there are a number of different languages

There are a number of different languages that had been differentiating for less than two thousand years.

Ardagastus said...

A split involving movement of farmers from Anatolia to Greece (which is simple, and we know that it happened) is a hell of a lot better explanation than the wholly conjectural "first split" of Indo-Europeans all the way around the Black Sea to arrive in Anatolia.


Except for space and time (which is actually what's to be proved by all these theories - where and when PIE was spoken), it is the same thing - a split.

Moreover, the oldest Anatolian languages (Hittite, Luwian) look like newcomers, not ancient survivors - little variety, not so many speakers (and those rather social and/or military elites), borrowings from few neighbouring languages (like Hattic).

There are a number of different languages that had been differentiating for less than two thousand years.

Between ~9000 and ~6000 BP there are some 3 thousands years. There's no consistent Latin vocabulary (wagons and wheels, agriculture, body parts, numerals, anything) preserved in most Romance languages without some significant changes, and even the cognates are pronounced differently. And that's a history of 2 millenia with written Latin as a quasi-permanent source of re-Latinization. How could some prehistoric languages survive 3 millenia virtually unchanged?

Dienekes said...

the oldest Anatolian languages (Hittite, Luwian) look like newcomers, not ancient survivors - little variety

Well, there is little variety in the IE languages spoken in Russia, but that doesn't stop you from postulating that Indo-Europeans originated there.

Or, the extreme homogeneity of modern Greek in the Balkan peninsula compared to the Slavic languages does not mean that Greek is younger in the peninsula than Slavic is.

Between ~9000 and ~6000 BP there are some 3 thousands years.

Once again Anatolian does not have wagon-related vocabulary.

As for the Latin languages, they diverged phonetically in part because they were spoken by different populations, and were imposed on them by elite dominance. Indo-European was spoken by the fairly homogeneous farming populations of the Balkan-Danubian area.

Maju said...

Well, Kurgans are completely absent from most of Europe (or Iran-India for that matter), so the steppe enthusiasts have a much bigger problem to tackle.

The burial style is only part of the matter: kurgans belong to a (long but obviously finite) period and vanish in the Bronze Age. But you can trace cultural continuity perfectly well. Kurgan in this sense is just a shorthand for a cultural continuum that we identify with Indoeuropeans. If there was any serious doubt on lack of continutity between, say, Corded Ware and Urnfields (who were not using tumular burials anymore), the whole theory would not stand.

But it stands very well.

And there are Kurgans in most of Europe, East of the Rhin, within that timeframe of Chalcolithic and early Bronze Age.

...so it is entirely possible that a handful of IE speakers from the Balkan-Danubian-Central European area mobilized the Kurgan populations and sent them eastward to Asia.

Only in the same sense that it's possible that they were from Mars. No other logic than "why not?", is ok for dreaming but not enough for science.

I'm thinking on it right now and I can't think of any European-specific migration or cultural flow eastward through the steppes between Gravettian and the Russian Empire. Only the Samara-Kurgan complex, assuming you consider the Samara valley to be in Europe.

Maju said...

guess my point about having a good leader is that it could just be luck.

That can explain one expansion (say Urnfields or Corded Ware or whatever) but not the whole chain of distinct expansions in space and time. In space between the Atlantic and the Indian Ocean, in time between c. 3500 BCE and present (in the Kurgan model) and between c. 7000 BCE and present (in the Anatolian model).

But the Cardial culture I think the origins are still a mystery...

Not so much. Cardium Pottery first appears in the context of early Thessalian Neolithic (but later vanishes from there, replaced by the painted one of Balcanic Neolithic - typically red and white, with some local variants). Then it is found in the Western Balcans (roughly classical Illyria: from Istria to coastal Albania and to the interior through Bosnia) and from there expands westward to Italy and Mediterranean France and Iberia (up to central Portugal and with some offshots in North Africa). They also expand into Biblos (sub-)culture (coastal Syria/Lbanon). There are instances of clear colonization but mostly it seems they assimilated natives as they expanded (Epipaleolithic local tool continuity).

So for all purposes we can consider CP to be of West Balcanic origin, distinct from the main Balcanic culture but with similar geographic origin in the Thessalian Neolithic. The main difference though may be economic: CP were clearly sea oriented and also did not use cattle much, instead focusing in sheeps/goats.

Its expansion westward is parallel to that of Balcano-Danubian Neolithic, just that following the coast and with more clear assimilation of natives. Neither one had much of a direct impact in the Atlantic (Portugal for CP, Mose-Seine for Danubian).

Dienekes said...

But you can trace cultural continuity perfectly well.

You can trace cultural continuity in most of Europe, without any evidence whatsoever (anthropological or archaeological) for the arrival of "steppe nomads".

And there are Kurgans in most of Europe

Most ignorant statement in this entire thread.

Only in the same sense that it's possible that they were from Mars. No other logic than "why not?"

We now know that lactase persistence started its growth in Central Europe/Balkans, and that the same allele encodes for it in Europe and South Asia. Clear evidence for involvement of European elements from far west of "Samara".

Ardagastus said...

Well, there is little variety in the IE languages spoken in Russia, but that doesn't stop you from postulating that Indo-Europeans originated there.

Or, the extreme homogeneity of modern Greek in the Balkan peninsula compared to the Slavic languages does not mean that Greek is younger in the peninsula than Slavic is.


My criteria are bound with logical "and"s, not with "or"s.
Greek's ancientness is proved by the borrowings it has from many languages (some of them ancient and extinct). The current Slavic spread is recent, and the main evidence is the relative lack of variety of Slavic languages contrasting with the large areas they occupy. And there was no IE continuity in the Russian steppes, as Russian is not the offshoot of some original IE dialect continuously spoken in the homeland.
Read here about the weaknesses of Anatolian hypothesis and of scholars considering Hittites intruders in their lands.

Once again Anatolian does not have wagon-related vocabulary.
Once again, Gray & Atkinson predicted that 6000 years ago PIE was split in several languages which will become later Tocharian, Greek, Armenian, Albanian, Balto-Slavic group, Indo-Iranian group, Celtic and a Germanic-Romance group. Each of these languages have a wagon/wheel vocabulary inherited from PIE and neither Gray & Atkinon, nor any other supporter of the Anatolian hypothesis could explain why. Even Renfrew had to modify his initial theory, adding a second homeland and split.

As for the Latin languages, they diverged phonetically in part because they were spoken by different populations, and were imposed on them by elite dominance. Indo-European was spoken by the fairly homogeneous farming populations of the Balkan-Danubian area.
The Balkan-Danubian material culture was not that homogenous. However, the language of any population (homogenous or not) varies, sometimes a lot. Even the modern Greek state with its strong nationalistic policies of ethnic and linguistic homogenisation and assimilation couldn't prevent the existence of diverging Greek dialects such as Tsakonian, or the unstoppable changes in Greek language. Modern Greek pronounciation is different from Koine which is different from Attic which is different from Mycenaean. And keep in mind that prehistoric languages didn't have such constraints, there was arguably more variation than in these languages we know today, preserved in written forms, learnt in schools, etc.

Maju said...

You can trace cultural continuity in most of Europe, without any evidence whatsoever (anthropological or archaeological) for the arrival of "steppe nomads".

You can't trace cultural continuity from Balcans/CE to the steppes (the opposite you can very well). And that is a central point that dismantles all the hypothesis (along with other stuff).

"And there are Kurgans in most of Europe"

Most ignorant statement in this entire thread
.

No. I know what I'm talking about.

We now know that lactase persistence started its growth in Central Europe/Balkans...

I'm not aware. I remind though that some aDNA study found that some fossil Danubians were all lacking that allele.

Lactase persistence so far has only been hypothetically linked to IEs but has no obvious IE-related distribution (much less Neolithic). Lactase persistence seems particularly high (almost fixated) among peoples that heavily relied on cattle of Atlantic Europe.

Read here about the weaknesses of Anatolian hypothesis and of scholars considering Hittites intruders in their lands

Exactly what I thought. Nice finding, thanks.

pconroy said...

Maju said:

Lactase persistence seems particularly high (almost fixated) among peoples that heavily relied on cattle of Atlantic Europe


If you are suggesting Atlantic Europe as the origin of LCT, then how would you explain the same allele in India - a back migration of Basques ;)

If you are not suggesting this, then the allele is intrusive and can only be linked to Neolithic (non-IE) invaders or later IE invaders.

Dienekes said...

I remind though that some aDNA study found that some fossil Danubians were all lacking that allele.

Duh, it was gene-culture co-evolution that caused the lactase persistence allele to increase in Central Europe. It's rather peculiar to expect to find the allele before the selection process even began.

Lactase persistence seems particularly high (almost fixated) among peoples that heavily relied on cattle of Atlantic Europe.

As the authors of the recent report make clear, areas of high frequency do not coincide with the area where the allele started to be selected. The latter happened in Central Europe, and the presence of the selected allele in South Asia represents a clear link with Central Europe and can't be explained by a "Samara"-only model.

Maju said...

If you are not suggesting this, then the allele is intrusive and can only be linked to Neolithic (non-IE) invaders or later IE invaders.

Its distribution matches mostly with economy, not culture or ancestry. I don't care where it originated: probably was floating around, the same as blond hair, and then became selected for among the milk lover peoples of the world.

It's not the kind of gene that tells much other than about the usage of drinking raw milk.

It's rather peculiar to expect to find the allele before the selection process even began.

I don't expect anything it's others who put way too much emphasis in a mere non-neutral allele.

As the authors of the recent report make clear, areas of high frequency do not coincide with the area where the allele started to be selected. The latter happened in Central Europe, and the presence of the selected allele in South Asia represents a clear link with Central Europe and can't be explained by a "Samara"-only model.

It's just a single allele highly favored by selection: a typical case of introgression: the gene moves and expands where it's selected for but regardless of any population move.

You cannot explain the high levels of lactose tolerance among Basques based on IE migrations nor in Danubian Neolithic either: the IE-speaking Basque neighbours are significatively more lactose intolerant and Danubian Neolithic never crossed the Garonne, even in their swan-song expansion. It's a random introgression between milk-lovers - whatever its origins.

Maju said...

More on this (rather irrelevant, IMO) issue of milk digestion ability. Take a look at Enattah 2007 for reference.

Enattah links not one but 9 different alleles associated to lactose tolerance and Basques as the most tolerant just after Utah whites (neo-Brits), followed by other non-IE groups: Finns and Fulani. Indoeuropean Iranians have exactly the same percentage (10%) as non-IE Arabs, their neighbours.

Jaakko Häkkinen said...

It is important here to separate Proto-Indo-European from Pre-Proto-Indo-European language stage. We know that Proto-Indo-European had a vocabulary related to wheeled vehicles, and the cognates are regular all over the language family (Dienekes, there are words for ’wheel’ and ’thill’ even in Anatolian branch). So the dispersal of Proto-Indo-European cannot have occurred as early as Renfrew’s model claims: it can have occurred only after 4000 BC. And Proto-Indo-European also had nature-connected vocabulary clearly pointing to the north of Pontic Sea, rather than south of it. (See Anthony, Mallory, Adams etc.)

Moreover, Proto-Indo-European had words common with both Uralic and Caucasian languages. The few Semitic words may well have been mediated by Caucasian languages, as they represent mostly cultural or technical innovations.

But it is still possible that the Pre-Proto-Indo-European was spoken somewhere south of the Pontic Sea – this cannot easily be disproven, nor proven. But it should be clear by now that Proto-Indo-European itself must have been spoken north of Black Sea. See Mallory 1989: In search of the Indo-Europeans for a comprehensive critique against Renfrew’s arguments.

The linguistic evidence is the only one that matters: there is no point claiming that all the carriers of certain archaeologically perceivable item or genetic lineage must have been spoken IE-language. Even if some genetic lineage or some cultural profile could be connected to IE languages in some place and time, the same lineage and culture could be connected to some other languages in some other place and time.

Dienekes said...

(Dienekes, there are words for ’wheel’ and ’thill’ even in Anatolian branch

Even Anthony and Ringe doubt this (pp. 64-65 of Anthony's book).

Kepler said...

Very OT:

Has anyone done deep clade test for J2 with Family Tree DNA?
What details you get? Just the clade, say J2a or whatever?
There is very little information on it on the site.
Thanks.

Anonymous said...

Quite a lively discussion,sadly mostly on PIE. A boring subject. Anything involving Europeans or Whites fair or dark is boring, obsessive and monotonous today. Semitic languages are more interesting due to their cultural and religious significance, though the Afro-Asiatic family invented by Greenberg is some what fanciful and imaginative.

That is the point of language studies. It is mostly based on very little facts and lots of assumptions. It is a bit like Astrology compared with Astronomy. Voodoo studies 101.

Today's Language groups seem to have their genesis back to 10 kya and seems to coincide with the invention of agriculture and animal husbandry. Those inventions increased populations dramatically, and changed a small group of primitive languages and made them meta language groups with accompanying linguistic changes. Without the invention of animal husbandry, the Kurgan builders would never have made the slightest impression in history. Just another insignificant group of primitives. That is what you have to remember with PIE or Proto Semitic, the barbarity of those people who spoke them.

Anonymous said...

I must add something.

Dienekes said that J2 is horizontal in its spread from its Middle Eastern homeland to the west and the east. J1 was said to be vertical in its spread.

That is fair enough for J2 an haplogroup introduced east and west from the Middle East by Middle Easterners of Neolithic vintage. But, J1 is more complex than that being found in higher frequencies, over 30%, in many places: Africa (Tunisia/Algeria/Ethiopia/Sudan), NE Caucasus, Middle East (Levantine corridor/Mesopotamia to the Gulf/South Arabia). Quite distant places. J1 also extends east west as does J2, with similar frequencies in Iran and Turkey either side of the classic Semitic or Arab zones and it decreases accordingly as it goes west into Europe, and east towards India. Its pattern in more like that of the Russian Orthodox Crucifix. That haplogroup's history is complex and much older than South Arabians like Yemenis or peripatetic Arabs like Bedouins.

It is correct that some J1 in North Africa, and most in the Sudan is of quite recent historic origin but most J1 in North Africa dates further back to the Neolithic age. I don't agree with Tofanelli when he states that most haplotypes of J1 are useless ethnically. There are overlaps, as races overlap, but ethnic trends came be seen

Andrew Lancaster said...

That can explain one expansion (say Urnfields or Corded Ware or whatever) but not the whole chain of distinct expansions in space and time. In space between the Atlantic and the Indian Ocean, in time between c. 3500 BCE and present (in the Kurgan model) and between c. 7000 BCE and present (in the Anatolian model).

Can you give details about what you mean concerning a chain of expansions?

Cardium Pottery first appears in the context of early Thessalian Neolithic (but later vanishes from there, replaced by the painted one of Balcanic Neolithic - typically red and white, with some local variants). Then it is found in the Western Balcans

I have never heard that before. Can you give a reference for this Thessalian Cardial? I know there have been attempts to say that there are similar types there, but I did not think there was much consensus on this? I've also heard that similar pottery is found in Byblos, but even this does not take us very far. Actually Cardial pottery is amongst the earliest pottery around the Mediterranean. (The Neolithic is of course older in Greece and the Levant, but not necessarily much different in terms of when pottery appears.) If you want something similar, and significantly earlier you have to go to the Sudan or NE Asia.

Maju said...

Can you give details about what you mean concerning a chain of expansions?-

Dunno. Whatever your model it's obvious that IEs did not expand through all their historical extension in a century or any sort of single wave. So it's not a matter of a leader or two but more probably a cultural and economical reality that kept the war machine going, even if intermitently.

Most other peoples simply did not do that and those who did mostly experienced a similar process of expansion similar to that of IEs (Semites, Turks, even Mongols but very late). They all were tribal, patriarchal and warlike and often found as seminomadic pastoralist societies.

I have never heard that before. Can you give a reference for this Thessalian Cardial?-

Can't find anything online. My notes from many years ago read:

"Otzaki (pre-Sesklo): origin of Cardial. Impressed pottery (punch, nails, fingers) or barbotine style, together with finer proto-Sesklo pottery". The notes were taken from a German book (in Spanish version) on European Neolithic/Chalcolithic, whose reference I did not keep.

I've also heard that similar pottery is found in Byblos, but even this does not take us very far.

Not really because it seems more recent in Lebanon than in the Balcans, and it is contextual of an older local culture: Amuq-Biblos, though only affects the Biblos facies. But we can't forget that the core of European Neolithic is not in Asia (even if it must be derived from there somehow) but in Greece and the rest of Balcans.

Actually Cardial pottery is amongst the earliest pottery around the Mediterranean. (The Neolithic is of course older in Greece and the Levant, but not necessarily much different in terms of when pottery appears.)-

I'm of the impression that Greek Neolithic pottery is among the oldest ones in West Eurasia. But West Asia is not as well combed archaeologically as Europe, so maybe we're missing something.

eurologist said...

@Maju:

I'm thinking on it right now and I can't think of any European-specific migration or cultural flow eastward through the steppes between Gravettian and the Russian Empire. Only the Samara-Kurgan complex, assuming you consider the Samara valley to be in Europe.

Most of nowadays Ukraine was unusually backwards for a couple of millennia when agriculture first prospered in Central Europe. Like first nowadays Poland, it suffered from the fact that initially there were no serials that could grow in the relatively short seasons, marked by excessive wetness both in spring and summer.

Instead, this vast region was initially predominantly sustained by hunter-gathers also practicing extensive fishing. Early agriculture before ~5,500 BC was rather primitive and limited to the extreme south, for the reasons described above.

Only just after ~5,500 BC the region prospered - after being overrun by agriculturalists from the northeast. They brought with them the serials and animals that worked - adapted to the climate, and with them the LBK long houses of the archeological record.

Guess what they also brought with them, long before Kurgan times...

At any rate, the IE they brought had 1,000 years to develop separately, forming the Satem branch, and later becoming the Slavic homeland.

eurologist said...

Clarification:

from the northeast

.. of central Europe, i.e., from the northwest from the Ukrainian viewpoint, i.e., going eastward.

Maju said...

Early agriculture before ~5,500 BC...

Erm? Before 5500 BCE? That is about the time when Neolithic reached Central Europe! Dniepr-Don culture (Neolithic) arises, probably on a local Epigravettian substrate, more or less when the Linear Pottery Culture is being formed in Hungary and when all Europe west of Budapest, north of the Carpathian mts. and Alps, was still Epipaleolithic. DD is older or at most contemporary with LBK (Danubian Neolithic).

Another thing is the SW Ukranian, but mostly Moldovan, Dniestr-Bug culture (also early Neolithic, 6th milennium), that would later be absorbed into the Danubian cultural area, as happened too with Wallachia and Bulgaria a little later. But even in that case, the Danubian is only a layer on an older layer and the Eastern Danubian of the Dniester-Maritza area (later the better known Chalcolithic cultures of Cucuteni-Tripolje and Karanovo-Gumelnita, probably the first European state of some size) shows many peculiarities that reflect their older Neolithic backgrounds.

Anyhow, this Eastern Danubian has almost nothing to do with Ukraine or the steppes: it just includes an area of SW Ukraine near Moldova. Most of the European steppes was Dniepr-Don culture, which is a distinct Neolithic culture that, sure, shows some Epipaleolithic reminiscences, and that would survive until the Sredny-Stog II stage, when the area becomes Kurganized. Before that, the area was origin of a NW migration or flow that reached as far as Scandinavia, but without replacing the pre-existent cultures where they were solid.

Anyhow, just to be sure: LBK (Danubian Neolithic) arose in the 6thmilennium in northern Hungary and spread mostly in the 5th milennium BCE. It did affect the Eastern Balcans (possibly as a reaction to the intrusion of Dimini-Vinca peoples, that seem to have destabilized the Balcanic Neolithic) but not directly Ukraine (Dniepr-Don).

Jaakko Häkkinen said...
This comment has been removed by the author.
Jaakko Häkkinen said...

New try:

Jaska:
>>(Dienekes, there are words for ’wheel’ and ’thill’ even in Anatolian branch

Dienekes:
>Even Anthony and Ringe doubt this (pp. 64-65 of Anthony's book).

There seems to be no reason to doubt, at least concerning the word *h2urg-i- > Hittite hurki ’wheel’ (regular cognates in Latin, Sanskrit and Old English).

Etymological Dictionary of the Hittite Inherited Lexicon
by Alwin Kloekhorst
https://openaccess.leidenuniv.nl/dspace/handle/1887/11996
(See page 424 for this ’wheel’)

eurologist said...

Maju,

You said: DD is older or at most contemporary with LBK (Danubian Neolithic).

What Dnieper-Don culture are you referring, to? The one I am familiar with was around 5,000 - 4,000 BC and started with hunter-gatherers and fishing, with only much later adoption of agriculture than the Danubian (LBK: 5,500 - 4,500). By ~5,000 - 4,000, LBK had already evolved and spread quite a bit north and east (incl. Lengyel derivative), due to locally-developed agricultural advances.

Sredny-Stog, the major agricultural and later horse domestication culture, wasn't even close to being contemporary (4,500 - 3,500).

Again, the reason is clear: agriculture in Poland, Moldovia, western and northern Ukraine required the climatic adaptions of serials and harvesting techniques (and suitable strains of cows and pig) that were developed with the LBK. That's why agriculture in the Ukrainian region was initially underdeveloped (including Bug-Dniester before LBK "soft invasion") and restricted to the very south.

In terms of population size, it is clear that the agriculturalists, initially coming from the west/northwest but as everywhere merging with locals, must have outnumbered the hunter-gatherers (of the wet and forested regions) by large proportions. There seems to be strong evidence that once Bug-Dniester essentially became a LBK drivative, it also attained a huge population size.

With regard to languages, I like the simplest explanations, best. Here you have a growing culture and population size in the same exact region that is generally viewed as the origin of later Slavic expansion. The simplest explanation is that the Slavic languages (and later SATEM groups) started out right then and there. Kurgans were just one, later facet of this, and IMO only but importantly involved with the eastward and southern spread of (SATEM)IE.

BTW, Anatolian agriculturalists spreading to the lower Balkans (and later advancing with the "southern agricultural package" to Italy and Spain) clearly were not IE, since there is no evidence of IE-related languages in the Mediterranean before peoples from the north entered - quite the contrary.

Maju said...

I can't find anything online and would need to go to the library to document this more exactly but I do have notes that confirm (at least to myself) my memory that DD and LP (LBK) beginning in the Early Neolithic (6th milennium in pan-European chronology). Actually this is true for DD but for LP we can only speak of prto-linear pottery in Northern Hungary in this period: the very genesis of the Danubian Neolithic, while in the case of DD the Neolithic with still major importance of hunting is more extended by the area already.

In any case DD is different in material culture from LP or Balcanic Neolithic: the pottery is different (large vases with neck in V shape) and burial is as well (collective burial in extended position with ochre, while Balcano-Danubians used the foetal position without ochre in individual burials with some offerings).

Sredny-Stog is not a culture, it is a site, whose phase I is fully in DD culture and whose phase II gives name to a mixed, confuse, transitional period between DD and Yamna. Sredny-Stog II lasts like 500 years (c. 3500-3000 BCE). It is irrelevant for this discussion but the claims of early horse domestication are related with DD, not Sredny-Stog II specifically (too late).

That's why agriculture in the Ukrainian region was initially underdeveloped (including Bug-Dniester before LBK "soft invasion") and restricted to the very south.

This is not that way AFAIK: in the Early Neolithic (6th milennium BCE) the Nord-Pontic region developed three local cultures: Bug-Dniestr, Dniepr-Don (very extrensive) and the small Surks-Dniepr culture (soon absorbed by DD). All them seem locally evolved as soon as the overall Neolithic know-how arrived. They are older than Danubian Neolithic properly speaking, though only DD will survive to later phases as such.

There seems to be strong evidence that once Bug-Dniester essentially became a LBK drivative, it also attained a huge population size.

There was a trend to grow in population everywhere as Neolithic consolidated, eventually leading to the increased social complexity and first clear hierarchies that characterize the Chalcolithic, but I don't think it has anything to do with this or that specific culture. DD itself also must have expanded demographically, and that's why they colonized the Baltic up to Sweden (Pitted Ware and clear influence in Funnelbeaker genesis). I think all cultures did one way or another.

Maju said...

With regard to languages, I like the simplest explanations, best. Here you have a growing culture and population size in the same exact region that is generally viewed as the origin of later Slavic expansion. The simplest explanation is that the Slavic languages (and later SATEM groups) started out right then and there. Kurgans were just one, later facet of this, and IMO only but importantly involved with the eastward and southern spread of (SATEM)IE.

Actually Satem is just an Indoiranian fashion that influenced proto-Balto-Slavics by sprachbund probably when the Scythians and Cimmerians ravaged the area. You can't ignore that Eastern IEs were once very influential in Eastern Europe: it's a recent well documented historical fact.

Slavic peoples surged in the 7th century CE out of nowhere (a swampy area between Ukraine and Poland) and expanded massively occupying some sort of vacuum. This is a very recent process and cannot be related with earlier processes of Indoeuropean expansion, at least not meaningfully.

BTW, Anatolian agriculturalists spreading to the lower Balkans (and later advancing with the "southern agricultural package" to Italy and Spain) clearly were not IE, since there is no evidence of IE-related languages in the Mediterranean before peoples from the north entered - quite the contrary.

In this I agree. But for me that (along with much other info) also means that surely the other parallel continental current (Balcano-Danubian) was not Indoeuropean either.

But, of course, one of the crucial problems is that Balcano-Danubian Neolithic never expanded into the steppes, so there is no way they could have initiated or even helped in any way the Kurgan phenomenon. Kurgans suffered some Danubian influence in the Balcans and Central Europe in the phase of their local consolidation, but not in Eastern Europe - not at all.

Dienekes said...

since there is no evidence of IE-related languages in the Mediterranean before peoples from the north entered - quite the contrary.

Lol, it so happens that the earliest evidence for IE languages is in the Mediterranean, in Anatolia and in Greece (and in Syria, if we include Mitanni). Not only that, but these languages were already extremely differentiated.

One has to wonder how Hittite, Mitanni Indo-Aryan, and Mycenaean Greek became so different from each other if they were all recent immigrants from the "north" at the time they enter the historical record.

pconroy said...

Dienekes,

As far as I can see the only way for IE languages to have an Anatolian origin and be distributed as they are, would be for an early spread via Eastern Anatolia and the Caucasus into the North Caucasus area around Maykop. Then you would need the Kurgan horse riders to abandon their own language and spread the language of the farming community they invaded.

This is of course possible, but is it probable??

Maju said...

One has to wonder how Hittite, Mitanni Indo-Aryan, and Mycenaean Greek became so different from each other if they were both recent immigrants from the "north" at the time they enter the historical record.

That's an excellent point, Dienekes.

But some 2000 years (time elapsed since the first known Kurgan expansion out of Samara) is enough to account for the differences between Romance, Slavic and Germanic languages. Plus I'd say that expansive languages always change more because of accelerated change caused by creolization. Plus differences between related languages and subfamilies seem to increase in a logarithmic fashion, not lineal. That's why we can hardly determine any traceable linguistic connection before some 10,000 years ago (est.)

Maju said...

Correcting myself:

But some 2000 years (time elapsed since the first known Kurgan expansion out of Samara) is enough to account for the differences between Romance, Slavic and Germanic languages.

I mean that 2000 years roughly (surely something more) between the Kurgan first expansions and the first signals of Hittite and Greek languages, same as from proto-Germanic (c. 2800), Latin (c. 1700) and proto-Slavic (c. 1300) to the various languages we know today. They would be, mutatis mutandi, like the Swedish and Indian English of their age, or like Haitian Creole and Timor Portuguese, I guess: you can still see that they have some stuff in common but they are totally unintelligible to each other.

eurologist said...

@ Dienekes:

Lol, it so happens that the earliest evidence for IE languages is in the Mediterranean, in Anatolia and in Greece (and in Syria, if we include Mitanni). Not only that, but these languages were already extremely differentiated.

No reason to be defensive. Written evidence is not evidence of language origin.

It's all a question of time frame.

Greece clearly was the major element propagating IE throughout the Mediterranean - first to Italy, then to Southern France and Spain. And later, from Italy (Rome)to non-IE Southern France pockets and Spain and Portugal.

However, that was much later. First, "from the North," Greece became IE long after Central Europe - which is pretty much undisputed. (My rough time line: 2500: Proto-Greek splits from IE in the very north Balkans; 2500 - 2000: Proto-Greek widely spoken in northern/central Balkans; 2000: Hellenic migrants enter Greece with Mycenean early Greek, etc.)

Hittites may have introduced IE into Anatolia 2000-1000BC, and people from the Caucasus may have played a role.

Dienekes said...

But some 2000 years (time elapsed since the first known Kurgan expansion out of Samara) is enough to account for the differences between Romance, Slavic and Germanic languages.

I didn't ask about "Romance", "Slavic", and "Germanic" languages, but about Greek, Indo-Aryan, and Hittite.

Also, 2000 years is not a very long time to account even for the smaller differences between the languages you mention. In the last 2000 years neither Germanic, nor Romance, nor Slavic languages have diverged nearly as much as these prehistoric languages are supposed to have diverged in a similar time period.

Dienekes said...

Written evidence is not evidence of language origin.

The point which you have missed is that Hittite and Greek (and Mitanni) appear first, and within centuries of each other in the archaeological record. The steppe hypothesis has difficulty in explaining how at so early a date, and so soon after the alleged breakup of PIE you already have three languages so different from each other. Why did these three languages become so different between 4,000-2,000 BC? There are many language families that have historical records in the last 2,000 years and none of them have diverged so much as these three languages are supposed to have in a similar time period -also, while supposedly being in a close proximity to each other.

Greece became IE long after Central Europe - which is pretty much undisputed.

Saying it is undisputed doesn't make it so. Central European farming cultures are derived from those of the Balkans, and we now know that the Central European farming population was certainly not particularly descended from the local hunter-gatherers. Therefore the hypothesis that they spoke languages that were descended from those of the Balkans at an earlier time is very strong.

Maju said...

Greece clearly was the major element propagating IE throughout the Mediterranean - first to Italy, then to Southern France and Spain.

If you mean Greek language, restricted to their colonial outposts (not too many in the far west), ok. But otherwise Indoeuropean in this part of Europe has Central European origins, the area where Western IE (sometimes called confusingly "European" - excludes Greek and Albanian, as well as other historical languages like Thracian, Dacian...) was spoken. The main vector of Indoeuropeization were Celts but in Italy it was Italics, also surely original from Central Europe. Rome (and not Greece) was the final blow of Indoeuropeizaton of the Western Med (replacing Etruscan, Iberian, Ligurian and possibly other remnant pre-IE languages).

Also, 2000 years is not a very long time to account even for the smaller differences between the languages you mention. In the last 2000 years neither Germanic, nor Romance, nor Slavic languages have diverged nearly as much as these prehistoric languages are supposed to have diverged in a similar time period.

Are you sure? I'm quite sure that an English speaker from India or Jamaica, faced with a Swedish text or speech would be unable to understand a word. Maybe educated polyglotes can see some similitudes after due examination (without going academic) but it's not too obvious. Just look at any random Swedish text and try to find words that look similar to English. I have one of those random texts now at another tab and in four short paragraphs I can discern the following words: "melodi", "dans", "musiken", "placerar" (pleasure? place?), "centrum", "personens", "folkmassan" (something related to "folk", I guess). You can see that all are loanwords of Latin or Romance origin (except "folk"). I'm not a native English speaker but I'm quite sure that even a native one would not find many more meaningful words.

In all honesty, I cannot find too many more words that are familiarly sounding in a random Romanian text either. And I am a native Spanish speaker and familiar with Italian since early childhood). And Romance has a significatively more recent divergence date.

You may be deluded by the relative conservatism of Greek language since the classical age (or maybe even since Mycenaean age) but that's because Greek has not expanded and assimilated new substrates, it has not gone any new process of creolization since its genesis in the 2nd milennium BCE, so all evolution is local drift and whatever loanwords it may have acquired by contact with other peoples. It has also been a literate (and liturgical) language since antiquity and that may hinder evolution somewhat as well.

I don't accept that languages evolve at constant rates. There is probably a basic truth in that but only if languages would diverge in ideally isolated previously "virgin" environments, so maybe as slow as Icelandic. But when languages expand over other substrate peoples, specially in lack of the modern high degree of interconnection (globalized media), they change fast specially at the beginning, when the natives switch languages and the foreign elite also makes linguistic concessions to the "popular" dialects, incorporating local words, accent and other modifications (including the usual simplification of grammar that happens in all creole languages). In brief: creolization accelerates linguistic change, as do cosmopolitan interactions, while continuity is a conservative force, only yielding to the unavoidable drift. So an expansive language will soon evolve faster than a language that remains for millennia isolated in some corner of the world.

eurologist said...

Maju,

Indoeuropean in this part of Europe has Central European origins, the area where Western IE (sometimes called confusingly "European" - excludes Greek and Albanian, as well as other historical languages like Thracian, Dacian...) was spoken. The main vector of Indoeuropeization were Celts but in Italy it was Italics, also surely original from Central Europe. Rome (and not Greece) was the final blow of Indoeuropeizaton of the Western Med (replacing Etruscan, Iberian, Ligurian and possibly other remnant pre-IE languages).

If you read my prior statement: (Anatolian agriculturalists spreading to the lower Balkans (and later advancing with the "southern agricultural package" to Italy and Spain) clearly were not IE, since there is no evidence of IE-related languages in the Mediterranean before peoples from the north entered - quite the contrary.
you will see that I agree with you 100%. Greeks and subsequently Romans were responsible for the final non-IE death nail in the Mediterranean - after IE surges from the North (that also formed Greek, IMO).

eurologist said...

BTW, Swedish is an extreme example, and may be older than most people think. Even Danish is tricky. On the other hand, Knowing German and English (and some rudimentary Plattdeutsch, I have no problem at all reading (which is easier) Dutch, and can understand about 50%-80% of it. Dutch, Northern German Dialects, and middle German/ Hochdeutsch as well as southern German have not changed much in about 1,000 years, and my guess is that based on localized early literature and their clearly present coloration, those dialects were already present 2,000 years ago.

The main changes in the last 2,000 years of Dutch/Frisian/German are the extinction of other (Eastern) dialects (that would sound really strange, today), and the second sound shift ~1,500 years ago. Not really all that much has changed since, and that includes more than 1,000 years with no or very little writing.

terryt said...

"One has to wonder how Hittite, Mitanni Indo-Aryan, and Mycenaean Greek became so different from each other if they were all recent immigrants from the "north" at the time they enter the historical record".

They were already differtentiated before they move south vias different routes? For example Mycenaean Greek to the west of the Black Sea, Mitannian to the east of the Caspian and Hittite through the Caucasus.

Dienekes said...

They were already differtentiated before they move south vias different routes? For example Mycenaean Greek to the west of the Black Sea, Mitannian to the east of the Caspian and Hittite through the Caucasus.

Right, but conside this:

1. In order to become differentiated they would have to occupy distinct geographical spaces. You would have to show some kind of linguistic isolation in the steppe region and adjacent areas that would be enough to create that level of difference, a formidable task.

2. Even if you could find such isolated northern homelands for the three languages, these would still be in relatively close proximity to each other (*) and share the basic postulated IE horse/military culture that they would later bring to their respective homelands. But, in 2,000 years, I don't really see the amount of linguistic change postulated for Greeks/Anatolians/Mitanni in any language family, even those whose members were much more geographically dispersed than the three languages would have been in their hypothetical northern homelands.

3. Finally, you would have to find three separate routes from the North to Greece, Anatolia, Syria. You would probably have to find a fourth route to account for Thracian and even a fifth one to account for Phrygian-Armenian.

4. Moreover, if we accept a gradual breakup of PIE since 4000BC, you will have to account for the gradual breakup of these languages in a span of 2000 years until they reach their final destinations. By implication, pairs of them must have been still united for some part of those 2000 years. So, you don't even have 2ky to explain away all the differences.

I would really love to see the geographical magic that would bring 3-5 very different languages from isolated homelands of the north to the south. Remember: you need to uncover isolated homelands in the steppe and adjacent regions to account for the enormous linguistic differences between Greek-Anatolian-Mitanni-Thracian. And, you need to find separate routes for all these languages to get them to their final homelands.

It looks like you'd need a traffic cop and miniatures of the Great Wall to explain away all this...

(*) Certainly not further away than modern Germanic, Romance, or Altaic languages that have not amassed nearly the same amount of linguistic divergence as Greek-Anatolian-Mitanni, despite being much more widely dispersed.

Maju said...

In order to become differentiated they would have to occupy distinct geographical spaces. You would have to show some kind of linguistic isolation in the steppe region and adjacent areas that would be enough to create that level of difference, a formidable task.

A formidable task indeed but IMHO:

1. The short-lived and complex Seredny-Stog II cultural area could well have been the first step of differentiation between Eastern (Samara) and Western (Seredny-Stog II) dialects. I'd say that in this case Greek would possibly belong to the Western branch but an expanded one, not the usual Western IE, that coalesced later in Central Europe (Baalberge to Corded Ware). Not all genealogical trees of IE place Greek so high in the tree.

2. There were also some Caucasian cultural areas (Maikop and I'm not sure about the Colchid culture) that look like a distinct branch (a second creole IE of the Caucasus), that may have given origin to the Hittite branch prior to its arrival to Anatolia.

3. Mittani Indo-Aryan obviously belongs to the third and more "genuine" IE branch, that remained behind, in the steppes, for longer: (proto-)Indo-Iranian.

The only languages I cannot explain (at least tentatively) are Albanian and Tocharian, that always show up as very old in the branching of IE. A possibility is that they might have segregated from the main group even before the Samara phase. Alternatively they may be so extremely creolized languages that this very process is what makes them appear higher ("older") in the graphs.

Finally, you would have to find three separate routes from the North to Greece, Anatolia, Syria. You would probably have to find a fourth route to account for Thracian and even a fifth one to account for Phrygian-Armenian.

I think we know those three routes (at least with some degree of certainty). The issue of Armenian, Phrygian and Thracian is very much disputed. In principle, Phrygians would have been a Thracian branch and Armenians (at least from classical sources), Phrygian colonists. But all this is quite confuse.

A serious possibility would be that Thracian would have coalesced in historical Thrace after the c. 3000 BCE invasions (Ezero culture, that does not show Kurgan burials but Dniepr-Don style ones - but arose after the Kurgan invasions anyhow and for many is ancestral to historical Thracians). proto-Greeks would then be a "sybling branch", maybe the one of Cotofeni culture (Kurgan) or maybe the one of Vucedol culture (the last Danubians with a IE elite, with megaron).

So I'd propose the following "rapid" branching out of IE:

PIE
> proto-Macro-Western IE (Seredny-Stog II)
>> proto-Western IE (Baalberge to Corded Ware)
>>> proto-Germanic (Single Burials)
>>> proto-Celtic (Altheim, Urnfields, western Hallstatt)
>>> proto-Italic (?)
>> proto-Thraco-Armenian (Ezero)
>> proto Greek (Cotofeni or Vucedol)
> proto-Middle IE (Maikop)
>> Hittite and related
> proto-Eastern IE (Yamna)
>> Indo-Iranian

As said before, not sure for Albanian and Tocharian but they should be early distinct branches too.

By implication, pairs of them must have been still united for some part of those 2000 years.

Western and Eastern (Indo-Iranian) IE did. The rest scattered very fast, it seems to me.

It looks like you'd need a traffic cop and miniatures of the Great Wall to explain away all this...

All you need is creolization in different places, as per above.

Maju said...

Sorry, I forgot about Balto-Slavic. Obviously is also within Western IE (Corded Ware).

Dienekes said...

Right, Greeks were in Romania and Croatia and Thracians were in Bulgaria, and they decided to "switch places" with the historical Greeks who were found south of the historical Thracians.

One also has to wonder what kind of horse-riding warriors would leave en masse the fertile plains of the Danube to settle in Greece of all places. And we have to think in terms of a real folk migration, because there was absolutely no trace of native Greeks or Proto-Greeks in Romania and Croatia in the historical period.

That's why I said you need a traffic cop to account for the final distribution.

BTW your model ignores the Illyrians, and the Tocharians.

Maju said...

The likely route Greece would be along the Vardar, right? It can be reached from Bulgaria but also from the Morava basin (Serbia). Both Vucedol and Cotofeni bordered (included partly in the case of Vucedol) this critical basin.

Why would they leave? For many reasons: didn't Slavs historically follow that same route, was not that route the one followed by Celts in their way to Anatolia? We can't discern to the fine details how this or that tribe, group of outlaws or highly organized conquest horde decided to go out and conquer but this is a pattern we see often among Indoeuropeans, so there's a cultural, ideological, factor "commanding" such process too. Bronze Age Aegean was a relatively rich area worth pillaging, I guess, and, as the conquest proved easy enough, they persisted on it until they conquered all the country.

The other possibility would be that Ezero people was the invaders and that Thracians invaded the country at a later date (from Ukraine). I am not sufficiently knowledgeable on the Late Bronze and Iron Age archaeology of the Balkans as to say for sure. Maybe even both Greeks and Thracians arrived together at a late date from the steppes. I leave that question open for someone more knowledgeable to explain. All I know is that I have read arguments in favor of the three possibilities.

Personally I favor Cotofeni because it's the culture that keeps using tumuli (kurgans) after the initial IE wave. And these tummuli are found c. 2000 BCE in northern Greece but this would fit too with a second "Greco-Thracian" wave from the steppes. I leave this issue open anyhow.

Dienekes said...

For many reasons: didn't Slavs historically follow that same route, was not that route the one followed by Celts in their way to Anatolia?

Both Slavs and Celts have left ample linguistic evidence of their presence north of Greece. But, where are the traces of the Proto-Greeks? Nowhere to be found. It seems they all packed their bags, called their travel agent, and headed for Greece, leaving absolutely no trace of their ever being in Croatia or Romania (or Bulgaria for that matter).

And these tummuli are found c. 2000 BCE in northern Greece but this would fit too with a second "Greco-Thracian" wave from the steppes. I leave this issue open anyhow.

There is absolutely zero evidence that Greek and Thracian are separated by a mere 500 years, since Greek is already archaeologically evident by 1,500BC and it is nothing like Thracian.

And these tummuli are found c. 2000 BCE in northern Greece

Where in northern Greece? And what kind of tumuli?

terryt said...

"In order to become differentiated they would have to occupy distinct geographical spaces".

To some extent they would have. If the proto-language had become widely spread through the steppe it would have formed a dialect chain. The separate languages would have developed from different regional dialects. The Indo- European language family is usually considered most closely related to Finno-Ugric (or whatever term is the correct one today) than to Semitic. Finno-Ugric is a northern language group. Semitic loan words are easily explained if we consider the possibility the various early Indo-European languages borrowed the same useful words from a single Semitic language that had managed to spread into parts of Anatolia and the Zagros. Perhaps Akkadian, or Eblaite, or a Western Semitic language.

"you would have to find three separate routes from the North to Greece, Anatolia, Syria. You would probably have to find a fourth route to account for Thracian and even a fifth one to account for Phrygian-Armenian".

The three routes are easily explained by my comment, 'Mycenaean Greek to the west of the Black Sea, Mitannian to the east of the Caspian and Hittite through the Caucasus'. Certainly the Mittanian gods were the same as the ancient Indian ones which suggests an Eastern Iranian connection. And I was under the impression that Greek was closely related to Phrygian, and probably to Armenian as well.

"Moreover, if we accept a gradual breakup of PIE since 4000BC".

I'm more than happy to accept a much earlier breakup. Which means the Kurgan people were responsible for just a later part of the expansion. Perhaps Thracian.

"Greek is already archaeologically evident by 1,500BC".

And it seems that Greek replaced an earlier language in the Agean at that time. Hittite's presence in Anatolia dates to about the same time, although many like to place it earlier to account for the anachronism of Abraham meeting Hittites. Something obviously led to the expansion and my bet would be the chariot's invention.

terryt said...

"BTW your model ignores the Illyrians, and the Tocharians".

Not mine though. The Tocharians simply originate from an eastern dialect, and I read somewhere that Illyrian was related to the Eastern Italian laguges, including Latin and ultimately Celtic. And I totally accept that creolisation will have served to differentiate the languages further.

Dienekes said...

Semitic loan words are easily explained if we consider the possibility the various early Indo-European languages borrowed the same useful words from a single Semitic language that had managed to spread into parts of Anatolia and the Zagros.

How could the "the various early IE languages" borrow the same useful words from a Semitic language that spread into Anatolia if they lived in Russia?

If the proto-language had become widely spread through the steppe it would have formed a dialect chain. The separate languages would have developed from different regional dialects.

We already know of two language families of the steppe region (Iranian and Altaic) who spent a few thousand years at least in the region, and were dispersed much more widely than the ancestors of Greek and Hittite could possibly have. Yet, there is no evidence that these language families have dialects anywhere near as different as Greek is from Hittite and Mitanni.

The three routes are easily explained by my comment

It depends what you mean by "easily". You don't only have to explain Greek as "West of the Black Sea", but also Thracian, Illyrian, Phrygo-Armenian. You'll have to explain how the Balkans (and Anatolia) ended up having numerous mutually unintelligible IE languages despite being geographically small regions with a fairly homogeneous Neolithic population.

You also have to explain how the entry of Anatolian languages from the resulted in them having almost zero historical presence in eastern Anatolia and being heavily concentrated in Central/Western Anatolia.

And it seems that Greek replaced an earlier language in the Agean at that time.

No, there is no evidence that Greek replaced any other language in the Aegean. The fact that Greek has non-Greek (some of which is IE, especially Anatolian, and some of which seems non-IE) vocabulary, does not necessarily mean that Greek replaced another language in the Aegean.

Going by that type of "logic", the fact that recent Greek has Turkish/Slavic/Albanian words in it, and there are plenty of Turkish/Slavic/Albanian toponyms in Greece, it means that Greek replaced Turkish/Slavic/Albanian in Greece.

Maju said...

Where in northern Greece? And what kind of tumuli? -

Seems my memory failed (my bookmarks got wiped out these last months and I had to search for my sources). It seems that the oldest known tumuli are in fact from Central and Southern Greece. They are nevertheless associated to rise of Mycenaean (properly Greek) culture and appear since the Middle Helladic (there's one as old as Early Helladic III) in relation with Minyan ware. Even the use of tholoi by Mycenaeans was differential, with single elite burials instead of the collective burials (a pre-IE signature) of Crete.

There is little time difference between the cultural reorganization of the eastern Balcans and Mid-Danub after the Kurgan-incursions and the rise of Mycenaean Greece. The Eastern Balcans were post-Kurgan since c. 3000 BCE, while the Middle Danub was only since c. 2400 BCE (contemporary with Corded Ware but distinct). Tumuli begin appearing in Greece since c. 2300 BCE.

Now, I don't know enough of the situation in the intermediate post-Vinca area to be sure what happened there in the 2400-2000 period.

This kind of issues, mostly coming from faulty, patchy, fragmented, research are anyhow similar to the caveats raised in the case of IA invasion in South Asia. I don't think this is any problem for Europeanists/Anatolianists to accept it (though it is for Indianists).

Maju said...

How could the "the various early IE languages" borrow the same useful words from a Semitic language that spread into Anatolia if they lived in Russia?

They are not Semitic words but actually from some other language(s) spoken by Neolithic peoples early on in West Asia. Both Semitic, IE and presumably other extinct languages borrowed them to varied extent as Neolithic spread (because they are Neolithic technical words and nothing else, right?).

Dienekes said...

Seems my memory failed (my bookmarks got wiped out these last months and I had to search for my sources).

I'll wait for your memory to return.

They are not Semitic words but actually from some other language(s) spoken by Neolithic peoples early on in West Asia.

Being Semitic in West Asia or non-Semitic in West Asia doesn't help you one bit in explaining how they found themselves to the "Proto-Indo-Europeans" north of the Black Sea.

Gioiello said...

Dienekes writes: “No, there is no evidence that Greek replaced any other language in the Aegean. The fact that Greek has non-Greek (some of which is IE, especially Anatolian, and some of which seems non-IE) vocabulary, does not necessarily mean that Greek replaced another language in the Aegean”.

When I studied Classics, I have always learnt that Ancient Greek had at least a 40% of its vocabulary not IE and I think it is difficult to demonstrate that Greeks were autochthonous.
That Pelasgians lived in Greece before Greeks I think it is difficult to deny and history and documents say when Greek tribes arrived there. The Linear A is probably linked to Luwian and only Linear B is Greek. You aren’t able to explain why the ancient Tetrapolis was named Uttenia and also Ilion presupposes Wilusa. The ancient GreeK “opuio” has no explanation in IE, but it has by Etruscan “puia” = wife. Etc. etc.

Dienekes said...

When I studied Classics, I have always learnt that Ancient Greek had at least a 40% of its vocabulary not IE and I think it is difficult to demonstrate that Greeks were autochthonous.

So what? Modern Greek has a lot of Italian/Albanian/Slavic/Turkish vocabulary, does that mean that Greeks came to a Greece inhabited by native Italians/Albanians/Slavs/Turks?

Nor should we forget that most of the non-IE vocabulary in Greek shows no parallels to any known non-IE language, hence it is better to say that it _cannot be reconstructed as IE_ rather than _it is not IE_. And, of course, it is expected that a fraction of the PIE vocabulary is not reconstructable in the languages close to the homeland (such as Greek and Anatolian) because it has been lost by drift in daughter languages.

The Linear A is probably linked to Luwian and only Linear B is Greek.

Indeed, Linear A can be quite plausibly be an IE language akin to Luwian. But, really, there is no evidence that Linear A represents the native language and Linear B an invasive language, just because Linear A tablets are older than Linear B. I'm sure that there are Greek inscriptions in Italy before there are any Latin ones, but obviously Latin is more ancient in Italy than Greek was.

I think it is difficult to deny and history and documents say when Greek tribes arrived there.

There are no such documents about any sort of migration of Greeks into Greece. There is plenty of evidence for migrations within Greece itself, and various stories about isolated mythical founders (from a diverse set of locations; no "coming of the Greeks" and no north-to-south arrivals)

As for Pelasgians, the ancient authors are divided into thinking them as either Greeks or as the progenitors of Greeks.

The ancient GreeK “opuio” has no explanation in IE, but it has by Etruscan “puia” = wife.

Not sure what word you mean, but anyway, even if there is evidence for Etruscan elements in Greek, what does that mean? Such elements could quite easily have been added to Greek, just as Etruscan elements were added to Italic languages after the migration of Etruscans' linguistic ancestors from Asia Minor.

Gioiello said...

Dienekes, we find Greek alphabet in Italy before every other simply because Greek invented it, transforming the Phoenician one, but certainly Greeks weren’t the first inhabitants of South Italy and we know when they arrived.
If we presuppose that Indo-Europeans were the farmers that escaped from the Black Sea submerged and gave life to the LBK in Central Europe, certainly the ancestors of Greeks were among them. Hittite, Luwians etc., Pelasgians, Rhaetian-Etruscans in Italy, were, perhaps, the most ancient branches of IE remained on the periphery of LBK. Certainly Greek language had some millennia of linkage with those central IE.
After probably they left Central Europe and migrated again to South. Probably they were linked with Armenians who reached the Caucasus. Ancient Greeks (the upper class) were fair haired, very different from to-days ones, and probably they came from North.
I don’t try in this explication anything dishonourable.

Maju said...

Being Semitic in West Asia or non-Semitic in West Asia doesn't help you one bit in explaining how they found themselves to the "Proto-Indo-Europeans" north of the Black Sea.

Neolithic, man. How do you say "internet" in Greek or Japanese or whatever? These things happen when there is technological innovation spreading fast. I would expect all Neolithic-specific words to originate in very few places regardless of linguistic affiliation. Of course locals can make up new words but the easiest thing is to copy what your more advanced neighbour already uses. That's how the word "cat" spread with the pet.

Samara culture was Neolithic too and was therefore somehow influenced by its West Asian core. Anyhow the Volga is not that far away from Kurdistan, is it? I have the impression that the Basque Country or Dravidian India are much more distant in fact and Neolithic words like the iri/ili/uli/uru/etc. for city arrived to both such extremely distant places anyhow.

It's just because the cultural wave (demic or not) spread from West Asia to everywhere around. And some words migrated with the objects and concepts they were meant for.

Dienekes said...

Dienekes, we find Greek alphabet in Italy before every other simply because Greek invented it, transforming the Phoenician one, but certainly Greeks weren’t the first inhabitants of South Italy and we know when they arrived.

And we find Linear A before Linear B, because the Cretans invented it and used it to write their language. Then Greeks adapted it, creating Linear B and writing their own language. We cannot infer that the language of Linear A pre-existed in Greece before the language of Linear B.

Ancient Greeks (the upper class) were fair haired, very different from to-days ones, and probably they came from North.

I won't take your word for it, I prefer to read the ancient sources.

Neolithic, man.

Precisely, the existence of Proto-Indo-Europeans in Asia Minor during the Neolithic easily explains the common terminology. The existence of Proto-Indo-Europeans of the steppe does not.

It's just because the cultural wave (demic or not) spread from West Asia to everywhere around.

Right, advocates of the steppe nomad hypothesis use the absence of agricultural terms in reconstructed PIE to argue against a role for the spread of farming from Anatolia, and at the same time have to find a role for the spread of farming from Anatolia to explain the Semitic parallels.

Maju said...

...there is no evidence that Linear A represents the native language and Linear B an invasive language, just because Linear A tablets are older than Linear B.

For Crete at least it does. Another issue is the mainland.

As for Pelasgians, the ancient authors are divided into thinking them as either Greeks or as the progenitors of Greeks.

Really? For Homer the Pelasgians are allies of Troy and among the peoples of Crete, speaking a distinct language ("language mixing with language side by side").

Herodotus is of the opinion that "Pelasgians ought to speak a barbarian language" and even claims that Athenians were themselves of Pelasgian roots, having lost their ancestral language.

There are way too many references for these non-Greeks (pre-Greeks) and to the quite conjectural nature of their non-Greek language to ignore them. There could have been other IEs but Strabo thinks they were the same stock as the Sicilian Aborigines (not the Sicels but the ones before them), who most likely were not. Of course, Strabo could be wrong, indeed.

Dienekes said...

For Crete at least it does. Another issue is the mainland.

Not even in Crete is this certain. But I was speaking about the Aegean area.

But, if you don't like my example, there are tons of better ones. There are Roman inscriptions before any Celtic ones in large parts of Europe. There is Slavic writing in most of the Balkans before there is any Albanian writing. And so on...

Really? For Homer the Pelasgians are allies of Troy and among the peoples of Crete, speaking a distinct language ("language mixing with language side by side").

Herodotus lists a whole bunch of Greeks who fought on the Persian side. When the Turks took Constantinople, there were more Greek Janissaries outside its walls attacking it than there were inside of it defending it. Being allies of the Trojans does not, of course, mean that they did not speak Greek.

Nor does Homer say anything about which of the languages of Crete were Greek or not, and the adjective he applies to the Pelasgians is dioi, directly linking them to the sky god of the supposedly invading Indo-Europeans whose origin was traced by the Greeks to Pelasgian Arcadia and "non-IE" Crete, the same Zeus whom he addresses as "Zeu Dodonaie Pelasgike" (Dodonian, Pelasgian Zeus).

Dienekes said...

Herodotus is of the opinion that "Pelasgians ought to speak a barbarian language" and even claims that Athenians were themselves of Pelasgian roots, having lost their ancestral language.

First of all, Herodotus' linguistic competence does not suffice to conclude that the Pelasgians did not speak Greek, and none of the examples of "Pelasgian" language he gives are from Greece itself.

And, Herodotus clearly indicates that the Hellenes were a part of the Pelasgians who split off from the main body and then grew in numbers by absorbing other nations of which the Pelasgian is the most important.

So, no, Herodotus furnishes no support for the idea of Greeks being invaders in a Greece occupied by Pelasgians.

Maju said...

Precisely, the existence of Proto-Indo-Europeans in Asia Minor during the Neolithic easily explains the common terminology. The existence of Proto-Indo-Europeans of the steppe does not

Not "easily". You (nor Renfrew) don't have any realistic model for how your "Indoeuropean Neolithic" expanded into Eastern Europe or the steppes.

As I said earlier, it would be easier to argue the Anatolian hypothesis in this aspect if the route to Samara and Ukraine would not be through Europe but through Central Asia. But guess that would clash with the implicit (and totally illusory) Eurocentrism in all this model.

Agricultural vocabulary could perfectly have crossed the Caucasus, the Black Sea, the Caspian Sea and/or the Central Asian oases together with the know-how of agriculture itself. In fact, it is quite likely that happened that way.

Right, advocates of the steppe nomad hypothesis use the absence of agricultural terms in reconstructed PIE to argue against a role for the spread of farming from Anatolia, and at the same time have to find a role for the spread of farming from Anatolia to explain the Semitic parallels.

Well, now you tell me which are precisely these terms. Because it can't be that they do share and do not share at the same time agricultural vocabulary, right? Maybe the devil is in the details?

terryt said...

"And we find Linear A before Linear B, because the Cretans invented it and used it to write their language. Then Greeks adapted it, creating Linear B and writing their own language".

The Cretans, an island people, are more likely to have the technology necessary to control trading routes in the Agean. The Greeks adapted the script and eventually invaded Crete, and presumably the wider Agean, once they had adopted the appropriate boating technology.

"We cannot infer that the language of Linear A pre-existed in Greece before the language of Linear B".

Without more evidence we certainly cannot assume the reverse either.

"We already know of two language families of the steppe region (Iranian and Altaic) who spent a few thousand years at least in the region, and were dispersed much more widely than the ancestors of Greek and Hittite could possibly have".

These have spread much more recently though. Altaic only in the last thousand years probably. And the Altaic languages Mongolian and Turkic are very different without considering dialects within each language group. Iranian's spread, too, is possibly just a little more ancient, and is certainly different from Sanskrit.

"advocates of the steppe nomad hypothesis use the absence of agricultural terms in reconstructed PIE to argue against a role for the spread of farming from Anatolia, and at the same time have to find a role for the spread of farming from Anatolia to explain the Semitic parallels".

I don't think anyone would argue with the idea that farming originated in or near southern Anatolia. But its spread from there need not have involved people speaking IE languages. Perhaps Semitic, as you suggest.

"How could the 'the various early IE languages' borrow the same useful words from a Semitic language that spread into Anatolia if they lived in Russia?"

Maju has summed it up, 'Agricultural vocabulary could perfectly have crossed the Caucasus, the Black Sea, the Caspian Sea and/or the Central Asian oases together with the know-how of agriculture itself. In fact, it is quite likely that happened that way'.

Dienekes said...

Not "easily". You (nor Renfrew) don't have any realistic model for how your "Indoeuropean Neolithic" expanded into Eastern Europe or the steppes.

There is absolutely no difficulty in that. We now know for example, that the pre-Neolithic population of Europe, including samples from "Samara", was dominated (if not exclusive) by mtDNA haplogroup U.

But prehistoric "Kurgan" South Siberians belonged to both U and various other West Eurasian mtDNA haplogroups. Similarly, upper-caste Hindus do have R1a1 at some frequency, but they also have J2a, which differentiates them from lower castes and tribals.

The archaeological scenario is not difficult to imagine, and I have already detailed it once:

You have Sakas and Tocharians moving into Asia, you have kurgans all the way to China, you have the clear evidence of the spread of technology from West-to-East.

Yet, what happened was that the steppe nomads who headed east sparked a change in lifestyle in the Altaic speakers who subsequently expanded west, wiping out IE speech from the steppe except for isolates. It is the language of the Altaics that won out, not IE.

The exact same thing happened on the other end of the steppe, when non-IE Kurgan groups interacted with the Indo-Europeans of the Balkans and Central Europe.

Without more evidence we certainly cannot assume the reverse either.

Ok, so you don't know that the language of Linear A preceded Greek in Greece or Crete, so you don't know that Greek was invasive in either Greece or Crete.

These have spread much more recently though.

Iranian languages have been spoken in the steppe for at least 3,500 years, probably more.

It is impossible to believe that in less than 2,000 years you'd see fragmentation of PIE into Greek, Indo-Iranian, Hittite precursor "dialects" in the steppe.

Why would a language family (IE) accumulate MORE linguistic variation in a smaller territory, in a shorter period of time, and despite the fact that it supposedly shared the same horse-based pastoral economy?

But its spread from there need not have involved people speaking IE languages. Perhaps Semitic, as you suggest.

There is absolutely no evidence to tie the spread of farming with "Semitic", as there are no Semitic remnants in either Greece or Anatolia where writing was known from very early on.

Maju has summed it up, 'Agricultural vocabulary could perfectly have crossed the Caucasus, the Black Sea, the Caspian Sea and/or the Central Asian oases together with the know-how of agriculture itself. In fact, it is quite likely that happened that way'.

Using that same logic, the so-called "wagon vocabulary" could just as easily have spread the other way, and there is no reason to tie the spread of Indo-European to that technological innovation.

Maju said...

I don't think that your "genetic evidence" is evidence of anything. U is a (consensually) very old haplogroup and probably arrived to Eastern Europe with the first (or second) colonists in the UP, J2a is found in such low frequencies among a particular endogamic caste that is most likely a founder effect that gives no evidence of anything but that founder effect as such. Furthermore, I think that part of the ancestry of Indian brahmins is in the IVC priestly class, not among IEs - for who priesthood was not particularly relevant: elsewhere the dominant class was the warrior class - only in India priests took over the hierarchy and that's surely because of pre-IE Chalcolithic substrate (in Sumer too the priestly class was dominant - it seems typical of early civilizations and too common in West Asia - cf. historical Hebrews, Egypt, etc.)

The archaeological scenario lacks a key evidence: migration to Eastern Europe. As I have pointed before there was NO migration or otherwise intense cultural influence from "peninsular" Europe (i.e. west of the Kaliningrad-Odessa line) to Eastern Europe in the Neolithic that could justify this direction for the spread of IE languages, culture and genes.

However Neolithic arrived to Eastern Europe it was surely by means of limited cultural influence and not mass migrations. If Samara people specifically were (partly) immigrants we don't know at the moment (though R1a1 points to a South/Central Asian origin) but it's quite clear that they did not arrive from elsewhere in Europe.

As I said before, a hypothesis of Anatolian Neolithic origins for IE would be much better off if it just equated IE and Neolithic altogether, allowing for this "IE" wave to reach South Asia via Iran early on and the steppes via Central Asia. Of course, the problem is that the historical linguistic evidence for West Asia is contrary to IE importance in the area before the late Bronze Age. All local languages but Hittite/Luwian (that seems intrusive) and Mittani IA (very clearly intrusive) are non-IE and each one of a different stock apparently. This situation is the same we find for Mediterranean Europe, curiously enough.

It is impossible to believe that in less than 2,000 years you'd see fragmentation of PIE into Greek, Indo-Iranian, Hittite precursor "dialects" in the steppe.

That's maybe your strongest point. I'd say that it is perfectly possible (creolization, like the already mentioned 40% substrate words in Greek) and we only see such differences exaggerated because most of the evidence for those languages is in fact from a later date (more in the 3000-4000 years range actually). Most of the oldest evidence for IE languages is of only 2000-2500 years ago (only some fragmentary stuff is from earlier times). We may be therefore missing some vocabulary/grammar links that have been blurred with yet another 2 milennia of evolution.

Using that same logic, the so-called "wagon vocabulary" could just as easily have spread the other way, and there is no reason to tie the spread of Indo-European to that technological innovation.

Fair enough. Actually Sumerians used wagons (at war) about the same time IEs did. While IE expansion may have carried some of these words, the invention is surely not IE-specific. IE phylogeny does not need of such specialized vocabulary to stand true in any case.

Dienekes said...

U is a (consensually) very old haplogroup and probably arrived to Eastern Europe with the first (or second) colonists in the UP

It's not important what "wave" of colonists brought it to Europe. What seems certain now is that populations from Central and Eastern Europe before the advent of farmers belonged (almost?) exclusively to it. We also know that both prehistoric Kurgan groups in the east and upper-caste Indians of today do not belong predominantly to this haplogroup, and they have some overlap with the available Neolithic samples.

So, your contention that "Samara Valley" Indo-Europeans were uninfluenced by Neolithic populations is false. Either they are not the ancestors of the eastern Indo-Europeans, or they are their ancestors but they were influenced by Neolithic Caucasoids from either the south or the west.

Furthermore, I think that part of the ancestry of Indian brahmins is in the IVC priestly class

The point is that J2a is socially stratified in Hindus; the wholly hypothetical absorption of an "IVC priestly caste" by Indo-Aryans really solves nothing, unless your model includes the slaughter of the IVC population and intermarriage only with the "IVC priestly caste". Otherwise, J2a (of the priestly caste) would be added to Brahmins, and J2a (of the regular IVC folk) would be added to the other castes, or would even be left outside the caste system.

we only see such differences exaggerated because most of the evidence for those languages is in fact from a later date (more in the 3000-4000 years range actually).

Not true. Hittite, Greek, and Mitanni are all evident about 2,000 years after the alleged breakup of "steppe Indo-Europeans". An interval of 3-4,000 years, combined with the earliest attested remains of Hittite of ~2,000BC leads to a breakup in 5-6,000BC which is definitely NOT what the steppe hypothesis holds.

Gioiello said...

(Gioiello) Dienekes, we find Greek alphabet in Italy before every other simply because Greek invented it, transforming the Phoenician one, but certainly Greeks weren’t the first inhabitants of South Italy and we know when they arrived.

(Dienekes) And we find Linear A before Linear B, because the Cretans invented it and used it to write their language. Then Greeks adapted it, creating Linear B and writing their own language. We cannot infer that the language of Linear A pre-existed in Greece before the language of Linear B.

You, Dienekes, who can read Aristotle in the original, which kind of syllogism is this?

Maju said...

It's not important what "wave" of colonists brought it to Europe. What seems certain now is that populations from Central and Eastern Europe before the advent of farmers belonged (almost?) exclusively to it. We also know that both prehistoric Kurgan groups in the east and upper-caste Indians of today do not belong predominantly to this haplogroup, and they have some overlap with the available Neolithic samples.

I don't understand your logic here, really. First, we're talking mtDNA, what is not likely to have been carried by invaders (mostly males, specially among the highly patriarchal IEs of old). Second, what happened with mtDNA H?

The point is that J2a is socially stratified in Hindus; the wholly hypothetical absorption of an "IVC priestly caste" by Indo-Aryans really solves nothing, unless your model includes the slaughter of the IVC population and intermarriage only with the "IVC priestly caste".

My model is open and flexible and ready to incorporate whatever new data (or not so new but that I may not know). But essentially I suspect that the priestly caste of India is not only of IE origin but partly of older origins. Ksatriyas (the warrior caste) actually often have much higher apportions of R1a1, or at least that is the impression I got from materials like Sahoo'06 and the like.

There could have been a "Cohen caste" in IVC and J2a be exclusive to it (or almost so). Personally I do think that, regardless of how distinct and unique is IVC, it had a significative West Asian influence in its formation, and Sumerian analogies in what has been identified as priests' sculptures have been suggested already by others more knowledgeable than me. Like in Europe, both agriculture and, later, civilization had an unavoidable West Asian reference at the time. For Europe this reference could be Troy or Syria/Cyprus but for India it must have been Sumer and Elam instead.

Not true. Hittite, Greek, and Mitanni are all evident about 2,000 years after the alleged breakup of "steppe Indo-Europeans"...

That's correct. What I say is that you are not comparing the scarce references we have of Mycenaean Greek and Mittani proto-IA but at least classical Greek and Sanskrit (1000 or more years older).

Anyhow, much of that diversification happened early on because of creolization. That is, I believe an important concept in the evolution of languages. For example, Castilian (Spanish) has only a few centuries but we can already detect several quite different dialects that, given the proper conditions, would easily diverge into very distinct languages. Maybe the most distinct dialect is not in Latin America but in the southern half of Spain, where it was spoken by people of different cultural background. How can Andalusian be so different from standard Castilian if it had only five centuries to evolve? Because it evolved "instantaneously" due to creolization. For English speakers, you can see the same process in the various pidgins (the one that calls more my attention is Indian English - the one spoken by the common people, not TV presenters). And this is happening in our highly globalized reality... imagine in the Bronze Age!

terryt said...

"But prehistoric 'Kurgan' South Siberians belonged to both U and various other West Eurasian mtDNA haplogroups".

But we know that languages can spread relatively independently of genes. Viz. your comment:

"what happened was that the steppe nomads who headed east sparked a change in lifestyle in the Altaic speakers who subsequently expanded west, wiping out IE speech from the steppe except for isolates. It is the language of the Altaics that won out, not IE".

But the older genes remain. Genes from further east do not carry near as far as the Altaic languages, or not in numbers anyway.

"Iranian languages have been spoken in the steppe for at least 3,500 years, probably more".

Actually 2500 years, probably more. Iranian may not have actually arrived in Iran until a little before the Medes developed. The Mittani may have spoken a language closer to Sanskrit rather than to Iranian. We do know that they mixed expensively with the previous inhabitants and their languages, as did the Hittites.

"Why would a language family (IE) accumulate MORE linguistic variation in a smaller territory".

It was probably not too small a territory. Proto-IE may have spread rapidly across the steppe from the Carpathians to the Altai before it broke into significantly different dialects, and began to move south. Anyway the problem of early diversity within a small territory is even greater if we confine IE's origin to Anatolia.

"For English speakers, you can see the same process in the various pidgins".

Not only pidgins. I had great difficulty understanding people, and being understood, the first few days I spent in the north of England.

Maju said...

Of course not just pidgins, Terry, but my point was that pidgins, creole languages diverge much faster because of the substrate influence, that alters the accent and pronunciation, introduces a large number of new words and expressions of aboriginal origin and may even alter the grammar, normally simplifying it but sometimes also complicating it.

Languages evolve fast but even faster if they do among "convert" speakers.

Dienekes said...

First, we're talking mtDNA, what is not likely to have been carried by invaders

This is a non-point, since we have both male and female burials from South Siberia which carry Caucasoid mtDNA.

my point was that pidgins, creole languages diverge much faster because of the substrate influence

It is a useless point, since Romance, Turkic, and Iranian spread over a much wider area, and experienced much more diverse substrate languages than Indo-European which according to the steppe model expanded into the fairly homogeneous "Old Europe" population of the Balkans.

There could have been a "Cohen caste" in IVC and J2a be exclusive to it (or almost so).

This is wild speculation.

It was probably not too small a territory. Proto-IE may have spread rapidly across the steppe from the Carpathians to the Altai before it broke into significantly different dialects

The point is that Hittite, Greek, and Mitanni are not descended from dialects from "the Carpathians to the Altai", unless you believe in a model according to which the precursors of Hittite, Greek, and Mitanni first dispersed all over the Eurasiatic steppe, and then converged within a thousand miles of each other.

Maju said...

I don't think that sparsely populated Siberia is the best reference for the rest of the planet but in any case, South Siberian mtDNA also includes Oriental lineages like F (from memory).

Romance is very very different from Latin. In fact all Romances are closer to each other than to their apparent parent language, classical Latin. Obviously creolization began as soon as Rome expanded, even in Italy, only that vulgar (creole) Latin got no prestige until the Middle Ages. Latin could well be considered an intermediate stage between archaic Western IE and Romances. And in fact it is also in between in the temporal dimension.

This is wild speculation.

Why? If Brahmins are modernly a hereditary caste and priestly castes are known to have existed historically elsewhere, specially in the Middle East, it just makes total sense that there was a "proto-Hindu" priestly caste before IA invasion, just that it got penetrated by IE lineages.

The point is that Hittite, Greek, and Mitanni are not descended from dialects from "the Carpathians to the Altai", unless you believe in a model according to which the precursors of Hittite, Greek, and Mitanni first dispersed all over the Eurasiatic steppe, and then converged within a thousand miles of each other.

That is exactly what happened, IMO:

·PIE > proto-Western IE (Ukraine) > Western IE (Central Europe) > Balto-Slavic, Germanic, Italic and Celtic
·PIE > proto-Western IE (Ukr.) > proto-Greek (SE Europe) > Greek
·PIE > Middle IE (Caucasus) > Hittite
·PIE > Eastern IE (Steppes) > Indo-Iranian > IA, Iranian and other

C. 3500-3000 there was quite a large expansion and diversification. You should count PIE as only existent before this date, then it's already dialectal forms (often creolized) evolving into the various branches. Since c. 3000 BCE IEs were already between the Elbe and Bulgaria by the west to wherever in the steppes they had already reached by the east. It's not anymore a local culture with an homogeneous society and language but a vast array of related ones, naturally including lots of various peoples who were not originally IE speakers but got engulfed in the surge, by grade or force (or something in between).

Dienekes said...

Why? If Brahmins are modernly a hereditary caste and priestly castes are known to have existed historically elsewhere, specially in the Middle East, it just makes total sense that there was a "proto-Hindu" priestly caste before IA invasion, just that it got penetrated by IE lineages.

As I already told you, this solves nothing, because you need to show why J2a would be concentrated in the "proto-Hindu" priestly caste. You are adding a hypothetical, and unnecessary level of complication for which there is no evidence, while there is plenty of evidence for a hereditary priestly caste among ancient Indo-Iranians (Brahmins among Indo-Aryans and Magi among Iranians).

That is exactly what happened, IMO:

Your model misses on all the Balkan languages besides Greek, and has no explanation for Tocharian.

C. 3500-3000 there was quite a large expansion and diversification.

No amount of "creolization" will create languages as different as Greek, Hittite, and Mitanni in a 1-1.5 millennia.

terryt said...

"Your model misses on all the Balkan languages besides Greek, and has no explanation for Tocharian".

For a start just add Tocharian to Maju's list:

·PIE > Northeastern IE (Steppes) > Tocharian.

Regarding the Balkan languages: I've always understood that Phrygian was simply a branch of Greek that crossed into Anatolia at some time (in fact someone can probably tell us when, I can't be bothered looking it up). Dacian, Thracian and Albanian may also be related to Greek, but more distantly.

And surely the fact that several of the language groups eventually converged by different routes towards the rich civilisation in Mesopotamia is not really surprising.

Dienekes said...

·PIE > Northeastern IE (Steppes) > Tocharian.

I don't see why Tocharian and Indo-Iranian, two languages so different from each other would evolve in the steppes in such a short period of time. If Tocharian was particularly close to Indo-Iranian, I could see a common origin to them both.

Dacian, Thracian and Albanian may also be related to Greek, but more distantly.

The existing evidence does not suggest a particular relationship of them to Greek, and even for Phrygian, the relationship is conjectural and not particularly close.

And surely the fact that several of the language groups eventually converged by different routes towards the rich civilisation in Mesopotamia is not really surprising.

I did not speak about "Mesopotamia", but about Greece, Anatolia, and Syria. You claimed that the splitting of Greek, Hittite, and Mitanni occurred in the wide dispersal in the steppe from the Carpathians to the Altai. So, why did languages that were dispersed from the Carpathians to the Altai end up within a thousand miles of each other in Greece, Anatolia, and Syria?

Maju said...

As I already told you, this solves nothing, because you need to show why J2a would be concentrated in the "proto-Hindu" priestly caste.

Well, the fact that it's found only (or almost only) in Brahmins and not, for instance Ksatriyas (or the many low caste people who also has presumably IE ancestry, as shown by R1a1), suggests it is not Indoeuropean but of some other origin.

Your model misses on all the Balkan languages besides Greek, and has no explanation for Tocharian.

Who has an explanation for Tocharian?! Or Albanian for the case.

Sincerely, I don't know how was the genesis of Tocharian. I'm just not sufficiently knowledgeable in Central Asian archaeology.

As for the rest, Albanian and Daco-Thracian. One theory is that they are the same thing. In any case, there were enough wildly different novel Balcanic cultures after the Kurgan raids period as to explain virtually anything. You have Ezero (culturally maybe more akin to Dniepr-Don than to Samara/Kurgan), you have Cotofeni (the more typically Kurgan of all for what I've read), you have this other culture of Wallachia and Dobrudja whose name I don't recall right now (that seems to have adopted cultural elements from Vinca) and you have Vucedol. Unlike what you see in Central Europe (quite homogeneous), the Balcans and Mid-Danub were then also "balcanized" with diverse post-Kurgan cultures. And if that's not enough, you have other Indoeuropeans, with their own complexity, just east of the Dniestr and also, soon after, east of the Bosphorus.

No amount of "creolization" will create languages as different as Greek, Hittite, and Mitanni in a 1-1.5 millennia.

AFAIK all we know of Mycenaean Greek, Hittite and the Mittani elite language is extremely little and highly fragmentary. Mittani looks like early Indoaryan, ok, but all we have is a handful of names.

Mycenaean Greek is the same: a handful of lists with names that are in many cases already substrate loanwords like "pasireu" (later "basileus"). Anyhow, how do you say "horse" in Greek? Hippos, right? Well, in Mycenaean times it was written down as EQU, rather suggesting Latin Equus. In fact all the animal words I see in the list of Linear B texts sound to Latin, rather than Greek (CERV = deer, EQU = horse, OVI = sheep, CAP = goat, SUS = pig, BOS = ox), though admittedly some may be way too similar in both languages.

Anecdotically, there is much more difference between Latin and, say, Spanish in most of these terms than between proto-Greek and Latin.

So let's be honest: the "Greek" of the Mycenaean period still was something else than classical Greek, even if some of the changes and adoption of substrate loanwords were already taking place.

You can't compare modern or classical Greek as if it was Mycenaean Greek. There are 1000 turbulent (and practically illiterate) years of difference. Mycenaean Greek was still PIE (probably Western PIE, very similar to that spoken in Central Europe) evolving to Greek, not finished Greek, as we know it.

Hittite is the only of those languages that actually has a minimally decent corpus and probably is the best window into a tentative approach to PIE or at least the middle branch of PIE as spoken around the Caucasus probably.

terryt said...

"why did languages that were dispersed from the Carpathians to the Altai end up within a thousand miles of each other in Greece, Anatolia, and Syria?"

Once the two-wheeled chariot had evolved the IEs were able to move south into already occupied regions. But they moved south via three different routes and just happened to all meet up again in Anatolia. Surely if they'd all started from there the languages would form a dialect chain rather than the at least three distinct groupings.

"If Tocharian was particularly close to Indo-Iranian, I could see a common origin to them both".

How close proto-Tocharian was to proto-Persian we have no way of knowing, but there individual places of origin could be very distant from each other. Anyway we only know a little of the various Tocharian dialects from just about 1000 years ago, the time when the two languages came into contact (again) in the Tarim Basin; Saka in the west of the basin and Tocharian in the east. The first 'Caucasian' mummies in the region date from about 3500 years ago. We've no idea if those people spoke proto-Tocharian or not but it's quite probable they did. The mummies date from around the time the Hittites first appear, the Mycenaeans, the Gutians, Kassites and Mitanni. Intersting, don't you think?

As for the date of dispersal of PIE. I strongly suspect that the early Indo-European expansion coincides with the domestication of the horse for purposes other than eating, probably about 6000 years ago. Although some of us seem to believe that modern humans lept onto a horse's back as soon as they'd crossed the Bab el Mandeb in their dugout canoes I'm not so sure. The horse came late to Mesopotamia. It is native to the steppe and is replaced below about 45 degrees by the onager, an animal difficult to domesticate. The domestic donkey is the African variety. In ancient Mesopotamia donkeys carried loads tied to their backs, as they do in many regions today. Four-wheeled wagons were usually pulled by cattle.

As soon as the people on the edge of the steppe realised they could use horses for those jobs loads and speed immediately increased, especially once someone invented the two-wheeled cart and the spoked wheel. They could now spread rapidly through what seems to have become a virtually unoccupied open steppe grassland. Humans had hunted mammoth on the steppe grassland in the Paleolithic but once the big game had became extinct humans had too.

Maju said...

Re. the Tocharians, someone explained to me long ago but I'm getting old and forget details. But that's why God created Wikipedia, right?

The case is that there is a Kurgan-like outlier culture in the Tocharian area (Afanasevo culture) that is as old as any other anywhere (except Samara, of course), dating to c. 3500 BCE. So it seems that while some IEs migrated westward, originating most European IE branches, others migrated to Altai and neighbouring areas, altering their culture in ways that suggest some sort of creolization too.

Meanwhile in the "central steppes", the "original" IEs, re-incarnated in the Yamna culture (and secondarily in the creolized Maikop culture of the Caucasus, leading to Hittites probably), probably developed the famous satemization and whatever other evolutionary changes into Indoiranian.

The details may be arguable but essentially the Kurgan model is pretty much proven.

Why did they migrate? To conquer and plunder, that's clear. From some moment on, this ideological item became important among these peoples (and was surely a way to get rid of excess young population that otherwise could prove troublesome). They were rough and rather poor but their neighbours were flourishing farming civilizations that had lots to offer in material terms to those who mananaged to subdue them. So they did. The chariot or the horse are incidental: Semites did exactly the same in West Asia with no vehicles.

terryt said...

"there is a Kurgan-like outlier culture in the Tocharian area (Afanasevo culture)".

That's the conclusion offered by the reference to the Tocharians I have in a book. Your assessment seems reasonable to me.

"The chariot or the horse are incidental: Semites did exactly the same in West Asia with no vehicles".

But the horse and horse-drawn vehicles allowed the Indo-Europeans to expand faster and further than the Semites were able to. Besides which the (probably) Semitic-speaking Hyksos had adopted the Chariot by the time they entered Egypt, so the technology spread beyond the genes or even the languages.

Dienekes said...

Once the two-wheeled chariot had evolved the IEs were able to move south into already occupied regions.

Let me get this straight: the "Indo-Europeans" drove all the way to the Altai and then drove back all the way to the Mediterranean.

Why didn't the steppe populations closer to Greece, Anatolia, and Syria "invade" these regions?

You either have to accept that:

(i) the ancestors of Greeks, Mitanni, and Anatolians all lived in the western part of the steppe, in which case their linguistic differentiation in the alotted time, and given the similarity in culture is impossible

or

(ii) that they executed an impossible maneuvre which spread them all over the steppe "from the Carpathians to the Altai", kept them separated for enough time (even though the examples of the Turks and Iranians shows that there just wasn't enough time), and then brought them back together within 1,000 miles of each other.

To say that this isn't parsimonious would be charitable...

The case is that there is a Kurgan-like outlier culture in the Tocharian area (Afanasevo culture) that is as old as any other anywhere (except Samara, of course), dating to c. 3500 BCE.

There is an interval of four thousand years between Afanasevo and the historical Tocharians. The vast majority of the terrain of the Eurasia changed their language within an interval of 4,000 years, so I see no reason to think that this wouldn't happen with respect to the Tocharians. Also, Afanasevo isn't in the "Tocharian" area, it is in the steppe to the north and west of it.

Also, if we look at the current population of the Tocharian area, there are at least three haplogroups represented in a high frequency: R1a, R1b, and J2, and others besides. The early sampled sites of Siberia are only R1a, so there is plenty of genetic room for the introduction of Western genes to the area.

They were rough and rather poor but their neighbours were flourishing farming civilizations that had lots to offer in material terms to those who mananaged to subdue them.

Right, they were "poor" in the fertile pasturelands and river valleys of the north, and they came to exploit the poverty of Greece's soil.

Maju said...

But the horse and horse-drawn vehicles allowed the Indo-Europeans to expand faster and further than the Semites were able to.

Surely. But early IE (and non-IE) carts were heavy 4-wheeled vehicles that had no military use and only limited transport improvement. The fact that the Kurgan lords were often buried with one, suggest that they were mostly a prestige item (just like in Sumer about the same time, where they used tamed onagres instead of horses and apparently were only used as royal vehicle).

Only in a later period (Bronze Age out of the Middle East, late Bronze Age there) did the 2-wheeled light chariot came to use. In the meantime, IEs had to travel and go to war either on foot or riding horses as some sort of primitive light cavalry or, more likely, mounted infantry. I suspect that horse riding, rather than the heavy and mostly useless primitive carts, was the element that gave speed and range to IE/Kurgan early migrants/raiders.

Besides which the (probably) Semitic-speaking Hyksos had adopted the Chariot by the time they entered Egypt, so the technology spread beyond the genes or even the languages.

Of course.

Maju said...

Let me get this straight: the "Indo-Europeans" drove all the way to the Altai and then drove back all the way to the Mediterranean.

No. They are different subgroups.

Why didn't the steppe populations closer to Greece, Anatolia, and Syria "invade" these regions? -

They did in fact. Let's see:

Tocharian branch: Some early Indoeuropeans went to Altai (Afanasevo) and eventually, in a mostly separate evolution, became the historical Tocharians.

Western branches: At about the same time other IEs were invading the Dniepr-Don basin (Serendy-Stog II) and probably founding that way the basic premises of European IEs (or most of them). A few centuries later some of these would invade or be summoned to upper East Germany, where they founded the core of what would be Western IEs (then Baalberge culture, 1000 years later Corded Ware). Still a little later, other IEs were plundering the Eastern Balcans, including Transylvania and parts of Hungary, leading at the beginning of the 3rd milennium to the creation of a series of novel cultures in the area that had probably IE dialects as their elite languages (Ezero, Cotofeni...).

Eastern/Middle branches: And about that same time, the Eastern IEs fully assimilated Ukraine (Yamna) and also incorporated parts of the Caucasus (Maikop, probably leading to Hittites).

All spawned from Samara culture, it seems, but they are different groups. In a few generations after divergence, they became distinct peoples. I even suspect that Western IEs may have helped the late Danubians of Baden culture and allies to fence off the Eastern IEs for a while in the Kiev area. But this was many centuries after their arrival and when the Danubians had become temporarily again the main power apparently. That's how Lubon culture, and its sucessors of Globular Amphorae and Corded Ware expanded into Eastern Europe (at the expense of the Eastern IEs of Yamna).

the ancestors of Greeks, Mitanni, and Anatolians all lived in the western part of the steppe, in which case their linguistic differentiation in the alotted time, and given the similarity in culture is impossible...

This is partly correct: the first division may well have been that between proto-Tocharians (Afanasevo) and all the rest (Samara and Seredny-Stog II). But by c. 3000, just 500 years after Afanesevo and Seredny-Stog II formed, each group was in a different region, many not anymore in the steppes (unless you consider East Germany, Poland or Bulgaria to be "the steppes").

Also, Afanasevo isn't in the "Tocharian" area, it is in the steppe to the north and west of it.

Actually it did include most of the habitable part of modern Sinkiang/Uyghuristan (i.e. the northern half, the south is mostly the vast emptiness of Taklamakan).

Also, if we look at the current population of the Tocharian area, there are at least three haplogroups represented in a high frequency: R1a, R1b, and J2, and others besides. The early sampled sites of Siberia are only R1a, so there is plenty of genetic room for the introduction of Western genes to the area.

Yes, that is interesting indeed.

Altai as such is nowadays anyhow mostly R1a (among these western haplogroups) so these differences might be (tentatively) due to localized founder effects. J2 should have been in Central Asia since at least Neolithic and it's possible that the same happened with the eastern branch of R1b (the most likely ancestral homeland for this lineage is West Asia, so they could have traveled together in the Neolithic or maybe even before). So guess that these lineages hitchiked the proto-Tocharian expansion.

Maju said...

Right, they were "poor" in the fertile pasturelands and river valleys of the north, and they came to exploit the poverty of Greece's soil.

They did not migrate as peasants, obviously. But anyhow the fertility of the northern (non-Mediterranean) lands is a "recent" phenomenon generated in the Middle Ages, with the introduction of technological improvements such as the heavy plough. Earlier (when only the so-called Roman plough was available) the Mediterranean was at least as fertile, if not more, and certainly much better connected and therefore rich in trade and all that stuff.

In the case of the Eastern Balcans, there was a rich Chalcolithic civilization worth to plunder. I can only imagine that the peoples of Karanovo-Gumelnita (probably the oldest centralized state in Europe) were just unprepared to meet the challenge of these raiders, who plundered them for one or two centuries until the civilization collapsed. Alternatively, as has happened in other better documented moments, it's possible that the increasing inequality of the advanced Chalcolithic society may have caused some sectors or classes to even welcome the barbarians as some sort of "liberators" at some point, as happened at times with the Germanic invaders of the late feudalizing Rome. We may never know for sure.

pconroy said...

Dienekes, Maju, Terry,

Take a look at the research Jean Manco has done on the subject of IE origins:

http://www.buildinghistory.org/distantpast/index.shtml

Here's what she has to say on the origin of the Hellenes (Greeks):

It was a time of vigorous change which ultimately gave rise to the Minoan civilization. There was increasing trade by ship across the Aegean. The Proto-Indo-European speakers of Usatovo had trade contacts with the Aegean world, to judge by some decidedly exotic artifacts in their possession. So the ancestors of the Greeks may have come by ship through the Bosporus into the Aegean regularly until some decided to settle on mainland Greece. As in Anatolia there was a lengthy period in which the horsemen seem subordinate to the farmers, before the former moved to seize hegemony of Crete around 1,400 BC. Again the new ruling class adapted the script of the old one for their language - an early form of Greek.95 Then came the Dark Ages of Greece, before the Greeks burst into history as a civilization that has influenced all subsequent ones in Europe. The expansion of the Greek world to southern Italy, Sicily and the European coast of the Black Sea seems to have left genetic traces in the spread of Y-DNA E1b1b1a2 (E-V13) and J2a1b1.

So she sees the genesis of the Hellenes as an amalgum of a Neolithic farming substrate and an Usatovo Culture elite.

pconroy said...

Other sources see the genesis of the Greeks in the Catacomb Culture of the Steppes:

http://en.wikipedia.org/wiki/Catacomb_culture


More recently, the Ukrainian archaeologist V. Kulbaka has argued that the Late Yamna cultures of ca. 3200-2800 BC, esp. the Budzhak, Starosilsk, and Novotitarovka groups, might represent the Greek-Armenian-"Aryan"(=Indo-Iranian) ancestors (Graeco-Aryan, Graeco-Armenian), and the Catacomb culture that of the "unified" (to ca. 2500 BC) and then "differentiated" Indo-Iranians.

Dienekes said...

But anyhow the fertility of the northern (non-Mediterranean) lands is a "recent" phenomenon generated in the Middle Ages, with the introduction of technological improvements such as the heavy plough.

Baloney. The Neolithic mega-cities of Cucuteni-Tripolye culture were not supported by a soil of poor fertility. Nothing on that scale existed in Neolithic Greece.

More recently, the Ukrainian archaeologist V. Kulbaka has argued that the Late Yamna cultures of ca. 3200-2800 BC, esp. the Budzhak, Starosilsk, and Novotitarovka groups, might represent the Greek-Armenian-"Aryan"(=Indo-Iranian) ancestors (Graeco-Aryan, Graeco-Armenian), and the Catacomb culture that of the "unified" (to ca. 2500 BC) and then "differentiated" Indo-Iranians.

Of course there is no evidence to tie "Late Yamna cultures" to Greeks (or to Armenians or even "Aryans").

Take a look at the research Jean Manco has done on the subject of IE origins:

...

So the ancestors of the Greeks may have come by ship through the Bosporus into the Aegean regularly until some decided to settle on mainland Greece. As in Anatolia there was a lengthy period in which the horsemen seem subordinate to the farmers, before the former moved to seize hegemony of Crete around 1,400 BC.


Horsemen seizing control of Crete? Obviously written by someone who hasn't set foot on Greece (or Crete).

"Regular sea trips from the Bosporus"? How imaginative, but not a shred of evidence in favor of the idea.

Also, what happened to all those "proto-Greek horsemen" who came via the Bosporus? Certainly no trace of them was left in "Usatovo". I guess they all decided to leave their cozy pasturelands in Romania, load up their horses in the non-existent long-range boats of the Bronze Age, in order to settle as "subordinates to farmers" for a few generations before conquering them in an unspecified manner.

Dienekes said...

I suspect that horse riding, rather than the heavy and mostly useless primitive carts, was the element that gave speed and range to IE/Kurgan early migrants/raiders.


Of course this is why we have so many early depictions of horse riders...

Maju said...

PConroy:

Isn't there a huge difference in time between the end of Usatovo (a clearly Indoeuropeanized Danubian culture of Westernmost Ukraine) and Yamna and the beginning of proto-Mycenaean remains. What happened between 3000 and 2000 BCE?

I can take the arrival by sea story but would need of a contemporary culture that was extant c. 2000 BCE, not a thousand years before.

Maju said...

Baloney. The Neolithic mega-cities of Cucuteni-Tripolye culture were not supported by a soil of poor fertility. Nothing on that scale existed in Neolithic Greece.

Sincerely I think that Cucuteni's story has been media-hyped. According to the same sources that says that their "cities" held like 10,000 people, it also says they burned their cities intently. This makes absolutely no sense to me.

I know of Cucuteni-Trypillia from some time now and never read such hyped stuff as that before. AFAIK, the neighbour Karanovo-Gumelnita culture of Bulgaria, Wallachia and neighbour districts of Greece and Turkey, was more the kind of "civilization center" of the area, though the inspiration for these Chalcolithic developments probably arrived from early Troy.

Whatever the case, it's clear that the great increase in fertility and productivity in the non-Mediterranean regions of Europe came only in the Middle Ages with these developments I mention: heavy plough and horse traction. This medieval agricultural revolution was the main pillar of the displacement of the center of civilization from the Mediterranean to Northern Europe, as the heavy plough is nearly useless (and possibly even counter-producing) in the thin Mediterranean soil.

But I agree that Greece was not in that period the center of anything: it was a relatively advanced culture (Rakhmani, evolved from Dimini) of the Bronze age but not central as would be later. Still surely worth a plunder and a conquest: if marginal East Germany or the arid Turkestan was, Greece surely was too. Dunno: some Vikings went to plunder France but others were happy to colonize cold Greenland, some Celts pillaged Etruria and Rome but others were content to conquer Galicia or Ireland. It depends of what opportunities are there for the militarized excedent population; often they just worked as mercenaries or probably intiated intra-ethnic wars as well. But eventually these energies were canalized against a neighbour.

Maju said...

Of course this is why we have so many early depictions of horse riders...

We don't (we don't have depictions of wagon drivers either, except in Sumer, where they were pulled by onagres, not horses) but we know that the Samara contemporary, Botai culture of West Khazakstan did ride horses. Also it has been argued that horse herding requires horse riding, otherwise there's no way to go as fast horse themselves go and they will run away. I am as sure as one can be that horse riding was already a common practice in the steppes prior to IE expansion, and obviously a vehicle for it.

Ardagastus said...

It is impossible to believe that in less than 2,000 years you'd see fragmentation of PIE into Greek, Indo-Iranian, Hittite precursor "dialects" in the steppe.

Check this paper by Andrew Garrett. Some illustrative excerpts:

In sum, especially if we allow that at least a few post-Proto-Greek changes must already have affected Mycenaean before its attestation (it is after all a Greek dialect), detailed analysis reduces the dossier of demonstrable and uniquely Proto-Greek innovations in phonology and inflectional morphology to nearly zero. Proto-Greek retained the basic NIE noun system, verb system, segment inventory, syllable structure, and arguably phonological word structure. In all these areas of linguistic structure, Greek was not yet Greek early in the second millennium. But if so, it hardly makes sense to reconstruct Proto-Greek as such: a coherent IE dialect, spoken by some IE speech community, ancestral to all the later Greek dialects. It is just as likely that Greek was formed by the coalescence of dialects that originally formed part of a continuum with other NIE dialects, including some that went on to participate in the formation of other IE branches. [...] It might not overstate the case to say that Mycenaean was a late NIE dialect with Greek vocabulary; a distinctively Greek phonological and inflectional profile was largely a development of post-Mycenaean history. (page 4)


Insofar as the formation of IE branches was a local process, and their characteristic innovations took place later than usually supposed, their phonological and morphological structures must have been closer in the centuries around 2000 BCE than has been thought. [...] The first-agriculturalists model posits a span of 3000-4000 years between PNIE and 2000 BCE. This means assuming two typologically incomparable periods, each three or four millennia long: a period marked by less phonological or inflectional change than is observed in any documented language, followed by a period when all IE languages were transformed by accumulating waves of phonological and morphological change. That is, the model requires the unscientific assumption that linguistic change in the period for which we have no direct evidence was radically different from change we can study directly. (page 7-8)


Dacian, Thracian and Albanian may also be related to Greek, but more distantly.

The existing evidence does not suggest a particular relationship of them to Greek, and even for Phrygian, the relationship is conjectural and not particularly close.


In these three articles a Romanian linguist suggests a stronger relationship between Thracian and Greek than commonly accepted.

Dienekes said...

Ardagastus, the point wasn't that Greek diverged a lot from PIE in 2,000 years, but that Greek, Hititte, and Mitanni had more differences between them than can be explained by 2,000 years of living apart.

Still, if Greek was fairly close to PIE as per your cited article, that would argue in favor of Greek being very close to the source of PIE rather than the opposite, since it is very difficult to envision how a PIE origin north of the Caspian, that underwent various transformations en route to Greece and a substantial substratum influence in Greece itself would remain simply an "IE dialect".

Ardagastus said...

Garrett's paper argues that all IE languages were fairly close in 2000 BCE, just like some dialects starting to diverge. He concludes PIE was spoken ~3500 BCE.
On page 7 you'll find also a table with five numerals in PNIE (PIE after Anatolian split), Proto-Greek, Proto-Indo-Iranian and Proto-Italic (the latter three around 2000 BCE) and they are quite similar.

I don't know where you get those "more differences [...] than can be explained by 2,000 years of living apart" from. As far as I know most linguists suggest a PIE split around 4500-3000 BCE (which means the differences between Anatolian and Myceneaen Greek are just as expected after 2-3 millenia of evolution), and it is mostly archaeologists and biologists who argue for an earlier split.

Dienekes said...

Garrett's paper argues that all IE languages were fairly close in 2000 BCE, just like some dialects starting to diverge.

That is nonsense, Hittite, Mitanni, and Mycenaean Greek are not "dialects" of any language, they are completely incomprehensible to each other. Not even Hittite and Luwian are "dialects", let alone Hittite and Greek.

As far as I know most linguists suggest a PIE split around 4500-3000 BCE (which means the differences between Anatolian and Myceneaen Greek are just as expected after 2-3 millenia of evolution)

I wonder how these linguists can say how much change can accumulate within 2 thousand years: they reject glottochronology and Bayesian phylogenetics when it comes to language change. So, all they have is their subjective opinions; I prefer quantitative evidence.

terryt said...

"the 'Indo-Europeans' drove all the way to the Altai and then drove back all the way to the Mediterranean".

The only relevance the Altai has to what happened in the Mediterranean is that people speaking variations on a single language spread from the Altai to the Mediterranean. And, obviously, languages at either end (and isolated within various regions, such as the separate ancestors of Mycenaean, Mittanian and Hittite) rapidly became very different.

"The early sampled sites of Siberia are only R1a".

Although we know that languages spread way beyond genes in many cases I strongly suspect that if any single haplogroup can be associated with an early IE expansion it is Y-hap R1a. I agree that the presence of J2a in upper caste Indians is interesting, but it has several explanations. Perhaps the caste system in India is older than the IEs, as Maju has alreay suggested, and social disruption with the IEs entry was not so great. Members of Y-hap J have a very good knack of convincing populations that they are special, especially when it come to being priests.

"and they came to exploit the poverty of Greece's soil".

They probably came to exploit the wealthy trading centres more like.

"Of course this is why we have so many early depictions of horse riders..."

I agree that Maju is mistaken in his support for ancient horse riding. It takes a brave man to jump onto a horse's back. A donkey perhaps less so, but they still buck pretty effectively.

"horse herding requires horse riding".

Not necessarily. It does require strong fences though.

"Hittite, Mitanni, and Mycenaean Greek are not 'dialects' of any language".

No, but the explanations of creolisation suggested by many here does adequately explain their rapid diversification. We can at least be certain that many of the original population in populated regions survived, whereas the IEs probably had the wider steppe to themselves for some time.

Ardagastus said...

That is nonsense, Hittite, Mitanni, and Mycenaean Greek are not "dialects" of any language, they are completely incomprehensible to each other. Not even Hittite and Luwian are "dialects", let alone Hittite and Greek.


"A language is a dialect with an army and a navy" some say. Language and dialect are conventional terms and there are dialects which are mutually incomprehensible (like some Italian dialects).
However Hittite and Myceneaen Greek (the language of Mitanni was largely a Hurrian language) were attested later than 2000 BC. Even so, they were not that different as you suggest them to be.

I wonder how these linguists can say how much change can accumulate within 2 thousand years: they reject glottochronology and Bayesian phylogenetics when it comes to language change. So, all they have is their subjective opinions; I prefer quantitative evidence.

There's no assumed rate of change (glottochronology) nor a flawed model behind a phylogenetic tree (like Gray and Atkinson's). Claiming a dubious Swadesh wordlist holds accurate evidence for linguistic evolution is also a subjective opinion, and worse, is mostly held by non-experts. Anatoli Fomenko also used "quantitative evidence" to prove that the entire chronology as we know it is wrong and that the Trojan war and the Crusades are the same conflict.

The differences between those proto-languages can compare with those between modern Romance languages. The changes from PIE to Proto-Greek can compare with those from Proto-Greek to Koine. There's no reason whatsoever why this prehistoric evolution should be almost three times slower than commonly imagined, that is an unscientific assumption as Garrett remarked. The main reason why Anatolian hypothesis and these methods of reconstruction are unpopular among linguists is that they run against the basic principles of historical linguistics, those principles used to theoretize a PIE language in the first place.

Maju said...

...it is mostly archaeologists and biologists who argue for an earlier split.

It rather surprises me that archaeologists would claim an earlier split (like supporting Renfrew's hypothesis). It surprises me because there is no archaeological evidence for any spread from the Balcans/Central Europe to the East and instead there are mountains of evidence for the spread of Kurgan peoples (as in Gimbutas' model).

I think it has more to do with how strongly one identifies with Indoeuropeans. Those who don't feel particularly attached to such identity, normally accept the much more solid Kurgan model, those who do, probably would prefer that Renfrew's model, in the case of Europeans, or the Indian origin model, in the case of South Asians would be true. But it's essentially wishful thinking: not wanting to accept the contradictions and violence of history.

Ardagastus said...

Here you'll find a text reconstructed for several IE languages in different stages of evolution. It supports a 3500 BCE split. Not all Indo-Europeanists would agree on those dates (the differences being however of 500-1000 years) or reconstructions, but I hope it makes it clear why a much slower rate of change is unlikely.

Ardagastus said...

It rather surprises me that archaeologists would claim an earlier split (like supporting Renfrew's hypothesis).
When I wrote "archaeologists" I mostly had Renfrew in mind, though obviously there are many supporting a north-Pontic homeland.

Dienekes said...

It surprises me because there is no archaeological evidence for any spread from the Balcans/Central Europe to the East and instead there are mountains of evidence for the spread of Kurgan peoples (as in Gimbutas' model).

There is much more evidence about the spread of the Neolithic than for any other population movement. Even die-hard anti-diffusionists like Renfrew accept that there was a movement of people across Eurasia during the Neolithic. Very few archaeologists see in the Kurgan migrations anything other than a local movement that affected parts of the Balkans and Central Europe, but did not really affect most of Europe.

But, we also have substantial genetic evidence for the inclusion of Neolithic elements in the eastward migrations across the steppe. For example mtDNA haplogroup N1a, which is the signature haplogroup of the LBK, and which differentiates them with the Paleolithic inhabitants of Europe (including those from Samara), was found in Iron Age Kazakhstan and modern Altaian Kazakhs, and upper caste Brahmins.

Maju said...

When I wrote "archaeologists" I mostly had Renfrew in mind, though obviously there are many supporting a north-Pontic homeland.

I'd like to know which are their logic. I generally see non-Kurgan hypothesis as amateurish and rather emotional than analytic. And in any case confronted with archaeological data. But if there is a substantial number of serious archaeologists supporting Renfrew's model, they should see some way of cultural spread from the Balcans to Samara valley, where the Kurgan and alternative models meet anyhow.

IMO this is the weakest point of all alternative models, be them Anatolian origin or Indus origin: not just that they still need Kurgans to explain the other half of the story but specially that they cannot propose a cultural connection between the alleged homelands and Samara culture (or Dniepr-Don either) on archaeological grounds.

Otherwise I wholeheartedly agree with you and thank you for your relevant and interesting input from the viewpoint of linguistics, Ardagastus.

Maju said...

There is much more evidence about the spread of the Neolithic than for any other population movement. Even die-hard anti-diffusionists like Renfrew accept that there was a movement of people across Eurasia during the Neolithic.

You cannot treat the Neolithic as a culture when it is obviously many, often unrelated or almost so. It is possible that there was a demic migration into the Balcans and maybe into Central Europe but this process can hardly connect with almost any other "Neolithic" anywhere. Even the Cardial Neolithic of the Mediterranean, possibly somehow connected with the Balcanic one early on at Thessaly (Otzaki), appears otherwise as wholly distinct; the alleged Anatolian connection with Hacilar is weak, not to mention the sometimes suggested Catalhoyuk (where are the lionesses and bulls in Balcanic Neolithic?); and there is no connection with Dniepr-Don either (for instance, these buried their people in the traditional Paleolithic style: extended position and ochre, while Balcano-Danubians used the foetal position instead - pottery styles also are totally different).

Very few archaeologists see in the Kurgan migrations anything other than a local movement that affected parts of the Balkans and Central Europe, but did not really affect most of Europe.

And don't forget Scandinavia and all non-Arctic Eastern Europe. In other words: all Europe except west of the North Sea-Rhine-Adriatic Sea line, with some areas of the SW Balcans spared at first but only for some time.

The areas not affected by Kurgan expansion west of that Rhine-etc. line, were later absorbed by Celts (Urnfields, Hallstatt, La Tène) and Italics (and Greeks in Italy too, sure), in processes that are not anymore "kurgan" (tumular) but that's only because cremation largely replaced inhumation in a well known localized evolution at that period of early Urnfields c. 1300 BCE. Anyhow, even in Urnfields culture there are still some kurgans.

But, we also have substantial genetic evidence for the inclusion of Neolithic elements in the eastward migrations across the steppe. For example mtDNA haplogroup N1a, which is the signature haplogroup of the LBK, and which differentiates them with the Paleolithic inhabitants of Europe (including those from Samara), was found in Iron Age Kazakhstan and modern Altaian Kazakhs, and upper caste Brahmins.

Very interesting but seemingly unrelated. It's a Neolithic signature original from West Asia, ok, but not apparently IE. Otherwise how would it arrive from the Balcans to India without affecting Russia? It's the same as with J2a that we discussed above.

Dienekes said...

Otherwise how would it arrive from the Balcans to India without affecting Russia?

What do you mean by "not affecting Russia"?

Dienekes said...

I strongly suspect that if any single haplogroup can be associated with an early IE expansion it is Y-hap R1a.

Unfortunately for that theory, the haplotypes of prehistoric steppe remains from cultures associated with the eastern spread of Indo-European show no matches in India or Iran, and even in Europe, the matches are very localized.

Ardagastus said...

I'd like to know which are their logic. I generally see non-Kurgan hypothesis as amateurish and rather emotional than analytic. And in any case confronted with archaeological data.
Besides Renfrew I recall Marek Zvelebil and Andrew Sherratt suggesting similar theories. If I'm not mistaken their focus is on a spread of language from Anatolia to Balkans contemporary with the spread of agriculture. Some arguments are for demic diffusion, some for language shift, consequently the interpretation of archaeological evidence is specific for each case.

Otherwise I wholeheartedly agree with you and thank you for your relevant and interesting input from the viewpoint of linguistics, Ardagastus.
You're most welcome.

Maju said...

Let's see, Dienekes: they do have matches with the original Baalberge/Corded complex area - quite precisely (for instance: lots of matches in East Germany but only one in West Germany). They also have matches in Anatolia but they have very few matches in the Balcans (all in the Vucedol area).

One could think of secondary redistribution via Turkic and Slavic migrations but it does not suggest a West to East migration: it suggests that the people who founded Afanasevo and those who founded Baalberge were "family". Maybe it was a single clan the one who pioneered the farthest early journeys of IEs: both Afanasevo and Baalberge share that they are very early and very remote settlements far away from the Russo-Ukrainian homeland.

It is interesting but says nothing in favor of a Balcanic homeland.

Maju said...

What do you mean by "not affecting Russia"? -

You and Renfrew propose that IEs spread somehow (no model other than generic appeals to "Neolithic") from the Balcans to the Steppes, where it became Kurgan (no explanation either about this radical alteration of the cultural baggage) and expanded into Central and South Asia, right?

So whatever item that arrived (within that hypothesis) from the Balcans to South Asia (or Altai/Turkestan) should have first coalesced for milennia in southern Russia, right? But all your supposed evidence, be it (Y-DNA) J2a or (mtDNA) N1a is not there.

So the most parsimonious explanation is that these arrived to South and Central Asia via some other route, like the Neolithic flows that, from West Asia, affected both Europe and Central/South Asia. IMO there are sufficient indications to suggest that the IVC priestly class was at least influenced by West Asian cultures (similar attire to Sumerian priests for instance) and that some of the genetics of that IVC priestly class (or caste) remains in modern Brahmins. The fact that IEs in most other cases had no priestly caste (Druidism was a pre-Celtic British invention, as per Roman sources) suggests that Brahmins, even if heavily permeated by the IE elite, also have some remainder ancestry from the pre-IE priests of IVC.

After all Hinduism existed in some form before the Vedic Age with all likelihood and, after the Vedic period, was partly restored, removing some of the IE Vedic influence (like the Vedic equivalent of Zeus/Jupiter: Indra, nowhere to be found nowadays).

Dienekes said...

So whatever item that arrived (within that hypothesis) from the Balcans to South Asia (or Altai/Turkestan) should have first coalesced for milennia in southern Russia, right? But all your supposed evidence, be it (Y-DNA) J2a or (mtDNA) N1a is not there.


What -except naive belief in genetic continuity across 6,000 years- makes you think that the current Slavic-Finnic population inhabiting Russia reflects the Bronze Age populations?

You and Renfrew propose that IEs spread somehow (no model other than generic appeals to "Neolithic") from the Balcans to the Steppes

There is no mystery about how the Indo-European populations from the Balkans spread to the steppe. Just as some of the Uralic/Altaic hunters spread to the steppe when they got horses, so did some of the IE populations from the Balkans and Central Europe.

terryt said...

"Unfortunately for that theory, the haplotypes of prehistoric steppe remains from cultures associated with the eastern spread of Indo-European show no matches in India or Iran, and even in Europe, the matches are very localized".

From your link: 'nearly all subjects belong to haplogroup R1a1-M17 which is thought to mark the eastward migration of the early Indo-Europeans'. We know that languages spread beyond genes so the fact there are no matches outside the central IE expansion is no problem.

"There is much more evidence about the spread of the Neolithic than for any other population movement".

And Y-hap J could easily be Neolithic, associated with the spread of agriculture. As you have pointed out the haplogroup is spread from the Balkans to the Northern Indian upper castes, through Northern Africa, Anatolia and the Zagros Mountains. That would certainly account for any postulated ancient relationship between the Semitic and Dravidian languages. In which case members of J hap could easily have spoken the language responsible for the shared Semitic-derived words adopted by the various IE dialects along agriculture's northern margin.

So how could Y-hap J remain in control of the upper castes in India? The theory is that warrior groups introducing a new language would have automatically taken over the top echelons of the existing society. It's possible they may have for a few generations, but to be of any use to an invader the infrastructure needs to be largely maintained. By the time the old guard becomes re-established the incoming males would be widely spread through the pre-existing population, possibly as a form of insurance for the original inhabitants.

terryt said...

"What -except naive belief in genetic continuity across 6,000 years-"

But hang on. Weren't you arguing just a little while ago, 'the haplotypes of prehistoric steppe remains from cultures associated with the eastern spread of Indo-European show no matches in India or Iran, and even in Europe, the matches are very localized'. Surely any ancient matches (I presume the comparison is between ancient South Siberians and modern Europeans, Indians and Iranians) would tend to become obliterated over time in this scenario too.

Dienekes said...

And Y-hap J could easily be Neolithic, associated with the spread of agriculture.

...

So how could Y-hap J remain in control of the upper castes in India


If J2a was "Neolithic" (and pre-Indo-Aryan) then it would be present at a higher frequency in the lower castes and at a lower frequency in the upper castes. The opposite is true, hence J2a is not Neolithic.

The only way around this would be to maintain that pre-Indo-Aryan society was already partitioned into a J2a-heavy and a J2a-light component, and that the Indo-Aryans mingled with the J2a-heavy component. This, is, of course, a completely conjectural idea without a shred of evidence in its favor.

From your link: 'nearly all subjects belong to haplogroup R1a1-M17 which is thought to mark the eastward migration of the early Indo-Europeans'.

I refer to the map of Y-STR matches which show no evidence of an Indian or Iranian connection. Remember that in the steppe model Andronovo is seen as ancestral to the Indo-Iranians.

Dienekes said...

"What -except naive belief in genetic continuity across 6,000 years-"

But hang on. Weren't you arguing just a little while ago, 'the haplotypes of prehistoric steppe remains from cultures associated with the eastern spread of Indo-European show no matches in India or Iran, and even in Europe, the matches are very localized'.


In India they've continuously spoken Indo-Aryan languages for at least 3-3.5ky. Moreover, the large population base and the caste system have almost certainly preserved the early Indo-Aryan population mix relatively intact (movement across caste lines occurred of course, but really little influence from outside India).

In Russia, by contrast, you have a plain which facilitates movement, and clear evidence for linguistic replacement and commingling of Slav and Finn in historical time.

So, no, I don't think it wise to think that the inhabitants of Russia 6,000 years ago are the same as today.

Nor is it a useful argument that N1a is not found in them, not only because Maju's statement about N1a's absence in Russians is false, but also because the evidence for the presence of N1a in LBK, Iron (and modern) Age Kazakhstan, and upper caste Brahmins builds a very nice case which has no real need for sampling any modern Russians to be supported. Indeed, if we sampled modern Central Europeans we would even think that N1a was absent in Neolithic Central Europe, which is false.

pconroy said...

Another thing missing in this comment thread is how, as elaborated by Anthony's book, groups of IE speaking horsemen from the Steppe probably were hired out as mercenaries or soldiers (or sometimes slave soldiers) by farming communities - largely to protect themselves from the incursions of other Steppe raiders.

We only need a cursory look at history to see how these systems can breakdown and lead to the protectors taking over. Examples that come to mind are:

1. Germanic speaking soldiers protecting the Roman frontier in Europe.
2. Vikings in Normandy protecting the rest of France from incursions from other Vikings.
3. Kipchak slave soldiers (aka Mamelukes) protecting Egypt
4. Gallowglass protecting Irish lords from attach

With this model, Steppe horsemen, can end up hundreds or thousands of miles from their homelands in a single generation and leapfrog barriers like sea and mountain.

pconroy said...

I should add to that list the Mitanni military elite among the Hurrians. These Mitanni were probably recruited from the Steppes, and then later took over their host population.

A similar story seems likely for the Hittites, taking over from their Hatti hosts.

pconroy said...

Here's more on the Hattians:

http://en.wikipedia.org/wiki/Hattians

including:

The Hattians and the Hittites apparently had different personal characteristics. Egyptian depictions of the Battle of Kadesh reportedly show long-nosed Hattian soldiers, while their Hittite leaders looked different according to the respected Turkish archaeologist, Ekrem Akurgal in a 2001 book.[6]
Akurgal claims that "The Hattians were still the great majority of the population in the Hittite period."[7] If true, the Hittite Indo-Europeans constituted a ruling elite within the Hittite Empire whereas the assimiliated Hattians were lower ranking members of Hittite society.

Dienekes said...

1. Germanic speaking soldiers protecting the Roman frontier in Europe.
2. Vikings in Normandy protecting the rest of France from incursions from other Vikings.
3. Kipchak slave soldiers (aka Mamelukes) protecting Egypt
4. Gallowglass protecting Irish lords from attach


In all these cases (not familiar with #4) the language of the larger population prevailed, not that of the "protectors".

pconroy said...

Here more info on the linguistics and PIE homeland:

http://www.hjholm.de/

Map:
http://www.hjholm.de/SLRD-map.jpg

Interactive Map:
http://www.hjholm.de/granim.gif

pconroy said...

On the origins of Greek, see Anthony pages 368, 369:

http://books.google.com/books/p/princeton?id=rOG5VcYxhiEC&lpg=PP1&pg=PA369#v=onepage&q=the%20origins%20of%20greek&f=false

Where he points the origin of Greek to the Catacomb Culture, and specifically the Ingul river area, between 2500 and 2000 BCE

Ardagastus said...

In all these cases (not familiar with #4) the language of the larger population prevailed, not that of the "protectors".
Actually if you look at a linguistic map of Europe, you'll see that Germanic languages are spoken south of Rhine and Danube in Western Europe, they are even spoken in today northern Italy.

However what about Hungary? This is a good example of some tribal steppe horsemen imposing their language over a relatively large territory.

pconroy said...

Here's a good link to maps of South East Europe, for all periods of history:

http://www.eliznik.org.uk/EastEurope/History/balkans-map/index.htm

Dienekes said...

On the origins of Greek, see Anthony pages 368, 369:

By his own admission he can't derive Greek from the steppes and doesn't even try to build a case. A couple pages of meaningless hand waiving.

pconroy said...

I interpret the data as he can derive Greek cultural practices from the Steppe, but can't account for how it got to historical Greece.

The answer as I've suggested previously is via mercenaries or "warrior bands", following trade routes to Greece.

Its interesting that some of the earliest Greek trade emporiums were founded on the Black Sea, especially the North shore - which would be exactly what one would expect, if the early Greeks had come from there in the first place. Like Americans trading with English etc.

Dienekes said...

Actually if you look at a linguistic map of Europe, you'll see that Germanic languages are spoken south of Rhine and Danube in Western Europe, they are even spoken in today northern Italy.

Germanic speakers took control of the entirety of Western Europe after the fall of the Roman Empire. They did not impose their language in the greatest part of the territory they controlled.

However what about Hungary? This is a good example of some tribal steppe horsemen imposing their language over a relatively large territory.

What about all the rest of the Balkans? The Slavs invaded the Balkans with Avar leaders, but it's the language of the Slavic farmers that prevailed, not that of the Avars? What about the Bulgars? Again, it was Slavic that prevailed, not the language of the Bulgars. What about the Franks, Visigoths, Longobards, Vandals, Normans, Kievan Rus, Indian Mughals , the Mongol empire, etc. etc.

The spread of language by elite dominance is the historical exception. In most documented cases in Europe, it is the language of the settled population that wins out. The only major exception is that of Romance languages, but here we have an advanced bureaucratic central state which advanced its language and culture methodically over many centuries.

Dienekes said...

I interpret the data as he can derive Greek cultural practices from the Steppe, but can't account for how it got to historical Greece.

In the context of the thousands pages of scholarship that has tried -without success- to trace the "coming of the Greeks" for more than a century, Anthony's two pages on the matter ranks as one of the least significant contributions.

Maju said...

What -except naive belief in genetic continuity across 6,000 years- makes you think that the current Slavic-Finnic population inhabiting Russia reflects the Bronze Age populations? -

And what about Altai that has suffered many more migrations from all directions but, curiously enough, preserves the ancient DNA as proven by your own sources?

There should be something.

There is no mystery about how the Indo-European populations from the Balkans spread to the steppe. Just as some of the Uralic/Altaic hunters spread to the steppe when they got horses, so did some of the IE populations from the Balkans and Central Europe.

There is: in this lengthy discussion that now has 166 posts you have been unable to propose a single cultural phenomenon that could reflect such highly hypothetical migration.

Not to mention that, unless you believe that horses were domesticated by Magdalenians in the UP, these beasts were domesticated in the steppe (probably Botai culture) and not the Balcans, where they are unknown before Kurgan invasions.

In Russia, by contrast, you have a plain which facilitates movement, and clear evidence for linguistic replacement and commingling of Slav and Finn in historical time.

Finnic peoples seem very old in the arctic and subarctic areas of Europe. They have surely been there since about the same time when IEs expanded, even if some are now assimilated as Russians.

As for Slavic linguistic replacement, I can only imagine that, like in the Balcans, it was just a layer on a substrate and that at least some genetic contuity should exist (if not the bulk). Normally peasants don't leave their villages, much less the region. And normally invaders don't want them to leave but to pay them taxes and swear loyalty.

And regardless your elitist viewpoint of history, the fact is that since Neolithic until deep into Industrialization, peasants have been everywhere like 90% of the people. Maybe they did not have an enviable existence but a viable one anyhow.

Dienekes said...

Its interesting that some of the earliest Greek trade emporiums were founded on the Black Sea, especially the North shore - which would be exactly what one would expect, if the early Greeks had come from there in the first place. Like Americans trading with English etc.

No, there were no early emporiums in the north shore of the Black Sea, nor any evidence that the Greeks encountered any long-lost relatives anywhere in the Black Sea when they colonized it. The north shore of the Black Sea was inhabited according to the Greeks by mysterious Cimmerians they knew nothing about, and later by Scythians who represented in all their cultural practices the very antithesis of a Greek.

Dienekes said...

And what about Altai that has suffered many more migrations from all directions but, curiously enough, preserves the ancient DNA as proven by your own sources?

Iron Age Altai.

Not to mention that, unless you believe that horses were domesticated by Magdalenians in the UP, these beasts were domesticated in the steppe (probably Botai culture) and not the Balcans, where they are unknown before Kurgan invasions.

Incorrect, there are horses in the Balkans before the "Kurgan invasions", they simply makeup a smaller part of the animal remains, suggesting that they were not economically as central.

And regardless your elitist viewpoint of history, the fact is that since Neolithic until deep into Industrialization, peasants have been everywhere like 90% of the people. Maybe they did not have an enviable existence but a viable one anyhow.

And peasant languages are everywhere dominant. Those invaders who "wanted them to pay taxes" don't care to make them speak their language, which is why in the vast majority of cases where a small number of foreign intruders have taken control of a larger agricultural population, it is the language of the natives that wins out.

Maju said...

@PConroy:

Another thing missing in this comment thread is how, as elaborated by Anthony's book, groups of IE speaking horsemen from the Steppe probably were hired out as mercenaries or soldiers (or sometimes slave soldiers) by farming communities - largely to protect themselves from the incursions of other Steppe raiders.

It is interesting that you raise this issue because I have for long suspected that the core Western IEs (Baalbarge culture, late 4th milennium BCE) could have been "invited" by the small Danubian culture of the mid-Elbe basin in a time that Western Danubians in particular seem to have been split and in intra-ethnic warfare (Michelsberg vs. Epi-Rössen). Just a guess but otherwise I don't see why or how such a group would have travelled the distance between the Dniepr (or Don or Volga) to the Elbe, bypassing other Danubian groups in between.

Another, already mentioned case, is maybe when Lubon culture (proto-Globular Amphorae, proto-Corded) took over the Kiev area, replacing a late Danubianized Dniepr-Don group. Again I suspect that the Danubians of Baden culture (the main Central European power of the time) eventually conceded that the best way to fence off an Indoeuropean attack was with an Indoeuropean defense.

It's only my speculation, of course, but maybe if these prehistorical incidents did not happen we would be speaking an Indoiranian dialect (or who knows?)

You may well be right about what the other IE tribes did in West Asia. We won't ever know for sure but it's not unreasonable at all. Plunder raids, full fledged invasions and mercenariate (eventually leading to coup d'etat) were probably different circumstantial aspects of the same process.

@Ardagastus:

However what about Hungary? This is a good example of some tribal steppe horsemen imposing their language over a relatively large territory.

An excellent example indeed.

Maju said...

Germanic speakers took control of the entirety of Western Europe after the fall of the Roman Empire. They did not impose their language in the greatest part of the territory they controlled.

The weight of literate and religious Latin was too heavy. A similar case can be seen in Mesopotamia, where Semites accepted Sumerian language (language of culture and religion) for long. Though in this case, Sumerian eventually receded and died off. A third case may be that of Mongol and Manchu China, where again the high prestige and literacy of Chinese language prevented that the invaders would impose their own. Yet another case is that of Persia after Islamic invasion, again we have a literary language of great prestige.

But these cases, even if rather well known, are surely exceptional: none of the peoples invaded by Indoeuropean tribes in late prehistory had a literary language nor an old prestigious political organization that could cause such conservative effects.

In China or post-Roman SW Europe, you could live in Chinese or Latin and even were tools that favored social mobility. I'm pretty sure that knowing "Danubian language" eventually was worthless in the IE-dominated societies of the 3rd milennium: you needed to speak what the king or nobles spoke: Indoeuropean.

And that's how the modern large languages have imposed themselves over other dialects or languages. You can live in a Basque-speaking village knowing only Spanish or French but you can't do business out of it, or almost anything, without speaking Spanish or French.

pconroy said...

Another example of Steppe warriors taking over an area, to which they were initially invited, would be in the Balkans, where the Serbs and Croats were both invited by the Byzantines (Emperor Heraclius) from the Steppes to establish themselves along the Northern border of Byzantine Europe empire, to protect them from attacks by another Steppe people, the Avars - of course they went on to establish their own kingdoms there, and imposed their language on the natives.

http://en.wikipedia.org/wiki/Serbs

Before their arrival on the Balkans, Serbs inhabited White Serbia, situated in present day Poland. The Serb settlement in the Balkans took place between 610 and 626 after being sent for by the Byzantine Emperor Heraclius to secure the Byzantine frontier from the problematic Avars.

Dienekes said...

The weight of literate and religious Latin was too heavy.

Romance languages are not descended from "literate and religious Latin" which is incomprehensible to Romance speakers, but from vulgar Latin. Indeed, "literate and religious Latin" was happily used by non-Latinate populations throughout Europe without affecting the popular speech.

But, why didn't the Avars, Bulgars, Kievan Rus impose their languages on the masses of Slavic peasants they ruled? The Slavs had neither literacy nor a substantial political organization.

Elite dominance is a real phenomenon but it requires a very organized political entity that can pursue a program of linguistic conversion over many generations. Arabic would not have spread in the Levant without the khalifate, nor Turkish in Anatolia without the Sultanate, nor Latin in Europe without the Roman state. There is no real evidence that "bands of warriors" can sustain -or would be interested in- this type of effort.

Dienekes said...

Another example of Steppe warriors taking over an area, to which they were initially invited, would be in the Balkans, where the Serbs and Croats were both invited by the Byzantines (Emperor Heraclius) from the Steppes to establish themselves along the Northern border of Byzantine Europe empire

Both "Serbs" and "Croats" were originally steppe Iranians who got intermingled with Slavs and gave them their name while losing their speech to that of the Slavic majority. The same happened with "Bulgarians" who were originally steppe Turks (or according to others Finns) who lost their speech to the Slavic majority. The same happened to the Rus who were Germanics who became Slavicized.

All these are good examples for how the language of the majority wins out, rather than the opposite.

Maju said...

But, why didn't the Avars, Bulgars, Kievan Rus impose their languages on the masses of Slavic peasants they ruled? The Slavs had neither literacy nor a substantial political organization.

I'd imagine that Avars, for the two centuries or so they ruled the Mid Danubian area, started to impose their language, indeed. Just that their rule was too short.

Kievan Rus did impose their language, early Russian, homogenizing what otherwise would surely be a conglomerate of diverse Slavic or non-Slavic tribes. Nobody speaks Pecheneg anymore, right? And Finnic peoples in the area are known to have been subdued by the Kievans. Not sure if you mean the Swedes that are sometimes claimed to be at the origin of the state, but in that case, it is controversial: probably both Swedish Varangians and local Slavic warriors co-organized that state.

The case of Balcanic Bulgaria is maybe the best example of this list but again Turks (Bulgars) and Slavs co-organized the new state. They don't speak Thracian or Greek anymore, do they?

Some examples of historical linguistic replacements by tribal invaders are: England (from Celto-Roman to Germanic), Middle and Upper Germania (from Celtic/Latin to Germanic), Hungary (from whatever to Hungarian), Slavic lands everywhere, notably in the Balcans, Turkey, most Arab-speaking countries, those countries of Latin America where genetics clearly place most of the population as more or less native (Bolivia, Peru, Guatemala, Mexico, etc.) and all Turkic speaking countries (Turks seem original from modern Mongolia in fact, though in some cases as Kazakhstan there might have been demic replacement, at least from the male side).

Other more proto-historical but rather well documented cases are: Hittite (and over them Phrygian) in Anatolia (over Hattic), Medic (proto-Kurdish probably) in Kurdistan and nearby areas (over Hurrian), Armenian in Armenia (over Urartean), Persian in Iran (over Elamite), Akkadian in Mesopotamia (over Sumerian)...

As you can see most of them include Indoeuropean peoples and languages or their "clones": Semites and Turks.

Maju said...

Elite dominance is a real phenomenon but it requires a very organized political entity that can pursue a program of linguistic conversion over many generations. Arabic would not have spread in the Levant without the khalifate, nor Turkish in Anatolia without the Sultanate, nor Latin in Europe without the Roman state.

Call it Caliphate or Khagante, I don't care. I'm quite sure though that in most cases these "bands of warriors" you speak or were pretty well organized, otherwise they would have gone nowhere. Was by chance Brennus, who vanquished Rome, less of a political and military leader than Quintus Fabius? was by chance Atilla less of a emperor than Theodosius? Their polities may have been weaker and less long lived often but enough, it seems, to plunder and conquer many others.

The case is that we do not nor probably can know how these "bands of warriors" or more likely tribal federations, empires or hordes, were organized. But if something we can learn from the examples of history is that they were sufficiently well organized to defeat those that were weak enough, no matter if these were centenary empires like Rome or tribes of peasants like the ones we can find mostly in early Chalcolithic Danubian Europe.

The question is that we cannot understand the political intricacies of those times based only on archaeology. We can speculate but little more. To have a realistic idea, we'd need a time machine.

Dienekes said...

Some examples of historical linguistic replacements by tribal invaders are

Your list is full of conjectural invasions, and examples of folk migration/demic diffusion rather than elite dominance.

The only worthwhile examples are of Arabs and Turks who spread because of their organization, i.e., they were organized states (Khalifate, various Turkish sultanates) which pursued a policy of conversion over many generations.

Of course nothing of the sort is attested for Europe; prehistoric Kurgan people almost certainly lacked the organization of the Arabs and Turks: they were illiterate, they formed small independent bands -no evidence of a great Kurgan "state" whatsoever.

The notion that the effected the linguistic conversion of the entire European continent is unbelievable.

Wave after wave of eastern invader (Finnic/Iranian/Turkic/Mongolian), most of them at a much higher level of social organization than the Kurgan people failed to change the language of even a small part of Europe (with the sole exception of the Hungarians), and yet we are supposed to believe that the much more primitive Kurgan people changed the language of the entire continent.

Credulity has its limits.

Dienekes said...

Was by chance Brennus, who vanquished Rome, less of a political and military leader than Quintus Fabius? was by chance Atilla less of a emperor than Theodosius?

Neither Brennus nor Attila made even a tiny change in a small part of the linguistic landscape of Europe, so I don't see why much more primitive Kurgan people would have changed the languages of the whole of Europe.

Ardagastus said...

Germanic speakers took control of the entirety of Western Europe after the fall of the Roman Empire. They did not impose their language in the greatest part of the territory they controlled.
They imposed in some of the territory (it may be that Frankish Austrasia is responsible for the spread of Germanic languages over Romance in Western Europe), which is good enough as it shows the language shift existed also in the case of early Medieval Germanic spread. Also Gothic was spoken in Crimea until 17-18th centuries or so.

What about all the rest of the Balkans? The Slavs invaded the Balkans with Avar leaders, but it's the language of the Slavic farmers that prevailed, not that of the Avars? What about the Bulgars? Again, it was Slavic that prevailed, not the language of the Bulgars. But the Slavic languages didn't spread by demic difussion as it's virtually impossible to have some tribes living in some Eastern European marshes (where traditionally the Slavic homeland is placed) to overcome demographically only in several centuries most of the Eastern Europe. Slavic languages spread also by language shift, and near the Roman border, where we have a better picture, there were more or less Romanized Gepids, Goths, Heruls, Celts, Sarmatians, Alans, Thracians, Illyrians, Greeks, Huns, Avars who learnt Slavic and became Slavs. Read this article for a recent synthesis of a theory of a Slavic lingua franca in the Avar khaganate.
And early Slavs were not only farmers, the sources attest them to be warriors, to have leaders of their own (Samo, Dauritas, Musocius, Perbundos, Peiragastus, Ardagastus), to be recruited as mercenaries, etc. Consequently there was also a Slavic military and political elite. However Hungarians are a better example for understanding IE spread because their honfoglalás matches better the scenario of horsemen conquering a land and imposing their language.

The spread of language by elite dominance is the historical exception. In most documented cases in Europe, it is the language of the settled population that wins out. The only major exception is that of Romance languages, but here we have an advanced bureaucratic central state which advanced its language and culture methodically over many centuries.
There are also linguistic substrata identified in most IE languages (also in Greek, see for example this paper) testifying for a new community of speakers changing the language of the original settlers (however not necessarily in the territory where that language is historically attested). Maju also provided several other relevant examples of language shift which should make us conclude safely enough that it is not an exception but a common phenomenon, possibly the most common one.

Ardagastus said...

I don't see why much more primitive Kurgan people would have changed the languages of the whole of Europe.
If one doesn't hold PCT or similar extremist views then obviously the IE speakers eventually changed the language of Europe.

Ardagastus said...

Of course nothing of the sort is attested for Europe; prehistoric Kurgan people almost certainly lacked the organization of the Arabs and Turks: they were illiterate, they formed small independent bands -no evidence of a great Kurgan "state" whatsoever.
Europe was mostly non-IE speaking in "Kurgan era" and most IE dialects were spoken in relatively small territories. Latin started out as the language of a city and its surroundings and swept over many languages both IE (Oscan, Umbrian, Illyrian, Celtic dialects etc.) and non-IE (Etruscan, Aquitanian, Punic) and probably also languages we'll never know they existed.

Dienekes said...

Latin started out as the language of a city and its surroundings and swept over many languages both IE (Oscan, Umbrian, Illyrian, Celtic dialects etc.) and non-IE (Etruscan, Aquitanian, Punic) and probably also languages we'll never know they existed.

Latin spread because it was the language of an organized state that pursued a policy of expansion for many centuries. Nothing of the sort can be postulated for prehistoric "Kurgan folk".

Moreover, Latin was the language of Central Italian farmers, not of steppe nomads: We do have countless names of nomadic people from the east who invaded Europe: Scythians, Avars, Huns, Bulgars, Alans, Mongols, to name but a few. None of them could so much as change the linguistic landscape of the settled populations of Europe in the least. Hungarians, a tiny bit of Europe is all that's left of all these steppe nomadic invaders.

Why should we believe that out of all the steppe groups that entered Europe only one (the Kurgans) not only affected the continent linguistically in a little way, but changed its entire linguistic landscape.

Ardagastus said...

Latin spread because it was the language of an organized state that pursued a policy of expansion for many centuries. Nothing of the sort can be postulated for prehistoric "Kurgan folk".
I don't see why would anyone want to do that, as virtually no one suggests that Proto-Italic, for instance, occupied the entire space from Atlantic to Middle East. Moreover, language families such as Celtic or Germanic experienced succesful expansions even though there was no prehistoric Celtic or Germanic state, but mostly tribal dominations.

Moreover, Latin was the language of Central Italian farmers, not of steppe nomads
They didn't spread their language because of superior farming technology, but through their military and political success, consequently the spread of IE languages with agriculture is rather unlikely as most (all?) "IE conquests" we know were political and military, not agrarian. If someone cultivates cereals better than you it doesn't look like a good reason to learn his language.

We do have countless names of nomadic people from the east who invaded Europe: Scythians, Avars, Huns, Bulgars, Alans, Mongols, to name but a few. None of them could so much as change the linguistic landscape of the settled populations of Europe in the least. Hungarians, a tiny bit of Europe is all that's left of all these steppe nomadic invaders.
Actually many of them did. Ancient Eastern Europe had an important group of Iranic speakers, Medieval Eastern Europe had Turkic speakers (Pechenegs, Cumans, Tartars - the latter surviving even today). All these nomads assimilated some local settlers and imposed their language for a while (they even left some durable traces such as toponymy), long enough for us to detect it as a historical change.
Also you're enumerating only steppe nomads, because otherwise Slavs, Celts and many other IE populations were also nomadic, though not in a "steppe way".

Why should we believe that out of all the steppe groups that entered Europe only one (the Kurgans) not only affected the continent linguistically in a little way, but changed its entire linguistic landscape.
Because this is the reality: most of today Europe speaks IE languages. There are only two options: 1) Europe was always mostly IE speaking (PCT) or 2) IE speakers changed drastically the linguistic map of Europe. Both Kurgan and Anatolian hypotheses support the second option.

Dienekes said...

Moreover, language families such as Celtic or Germanic experienced succesful expansions even though there was no prehistoric Celtic or Germanic state, but mostly tribal dominations.

The spread of Celtic, Germanic (and Slavic) was primarily one of folk migration or demic diffusion NOT of elite dominance. We see this clearly in the states where there was indeed clearly a Germanic elite over a foreign population (among Latins, Slavs, and Finns alike) where the language of the population remained unchanged.

Moreover, both Celts and Germans were not steppe pastoralists but European farmers/stockbreeders. So, where are all the examples of steppe warriors changing the language of Europe? There have been numerous examples, and only the Hungarians succeeded and for a tiny bit of Europe.

:: Why should we believe that out of all the steppe groups that entered Europe only one (the Kurgans) not only affected the continent linguistically in a little way, but changed its entire linguistic landscape.
Because this is the reality: most of today Europe speaks IE languages. There are only two options: 1) Europe was always mostly IE speaking (PCT) or 2) IE speakers changed drastically the linguistic map of Europe. Both Kurgan and Anatolian hypotheses support the second option.


The Kurgan hypothesis does NOT support the second option as it was not a pan-European phenomenon, but one that affected only the periphery of Europe, moreover none of the known steppe intruders had any success in changing the language of Europe.

The spread of farming -unlike the Kurgan intrusions- is clearly one process that affected most of Europe.

It's up to those who think that the large agricultural populations of Bronze Age group changed completely their language into IE due to "steppe pastoralists" to explain why the later agricultural populations of Europe remained completely unaffected by repeated intrusions by real centrally organized eastern cavalry armies.

Ardagastus said...

The spread of Celtic, Germanic (and Slavic) was primarily one of folk migration or demic diffusion NOT of elite dominance. We see this clearly in the states where there was indeed clearly a Germanic elite over a foreign population (among Latins, Slavs, and Finns alike) where the language of the population remained unchanged.

There's no archaeological or anthropological evidence of overwhelming migrations in these cases (and the historical evidence is often dubious and contested, see W. Goffart for Germanic migrations, F. Curta and P. Barford for Slavic migrations, etc.), on the contary most evidence supports language shift (through some sort of elite dominance). Celts expanded from British Isles and Iberian peninsula to Anatolia (Galatians) being rather military elites (the case is quite obvious for Balkans and Anatolia where they are mentioned as such by historical records), and Germans also expanded over Celts and other non-Germanic populations also as military elites. Witnessing these wide linguistic spreads of Celtic, Slavic and even Germanic (from far north to Alps, from Britain to Black Sea) demic difussion becomes unverosimile as we can't reasonably expect that repeatedly in European history the population of a small territory overcame demographically vast areas of Europe. There are regions of Europe changing their dominant language several times in history! Where is the evidence for some flooding migrations?

Moreover, both Celts and Germans were not steppe pastoralists but European farmers/stockbreeders. So, where are all the examples of steppe warriors changing the language of Europe? There have been numerous examples, and only the Hungarians succeeded and for a tiny bit of Europe.
Steppe pastoralists were stock breeders. Also, contrary to your claims, there were already several examples like Scythians, Magyars, Tartars who succeeded to impose their language over some Eastern and Central European territories with little if any resistence. And there's no theoretical constraint that initially the IE languages spread over huge areas (Slavic languages notably started to expand only some 1500 years ago), so there's no reason to expect a more exceptional situation than those we already know for other steppe horsemen.

The Kurgan hypothesis does NOT support the second option as it was not a pan-European phenomenon, but one that affected only the periphery of Europe, moreover none of the known steppe intruders had any success in changing the language of Europe.

But actually it does, as Anthony, Mallory et al. point out. The IE speakers changed the linguistic map of Europe during several millenia, not in their initial expansion. They came as horsemen but they eventually succeeded through some other ways but horseriding and burial rites.

The spread of farming -unlike the Kurgan intrusions- is clearly one process that affected most of Europe.

Most of Europe started to learn IE languages relatively recently (in 1st millenium BCE and after). Do you really believe Western Europe switched from Paleo-Iberian and Aquitanian to Celtic and Latin because of farming?

It's up to those who think that the large agricultural populations of Bronze Age group changed completely their language into IE due to "steppe pastoralists" to explain why the later agricultural populations of Europe remained completely unaffected by repeated intrusions by real centrally organized eastern cavalry armies.
But they didn't remain "completely unaffected" therefore there's nothing to explain. And the burden of proof is on the supporters of demic difussion, as they claim that the most common scenario of linguistic spread is a demographic expansion in an underpopulated space, possibly with the extermination/migration of the previous speakers. Where is the evidence? Language shift may leave no trace in the archaeological or anthropological record, such demographic changes certainly would.

Dienekes said...

Celts expanded from British Isles and Iberian peninsula to Anatolia (Galatians) being rather military elites (the case is quite obvious for Balkans and Anatolia where they are mentioned as such by historical records)

This is an example of a folk migration, NOT of elite dominance, as there is no evidence that the Celts in either the Balkans or in Anatolia imposed their language on the natives at all, and indeed their language was lost very quickly by that of the larger surrounding population.

There are regions of Europe changing their dominant language several times in history!

Which makes it even more surprising that absolutely none of the language shifts within Europe (with the exception of Hungarian) originated in the steppe region, the alleged homeland of the PIE. Quite simply, steppe nomads, despite their repeated intrusions into Europe have been singularly unsuccessful in effecting linguistic change in Europe.

But they didn't remain "completely unaffected" therefore there's nothing to explain

They did. We have repeated movements of Iranian, Turkic, Mongolian, and Uralic speakers into Europe from the Eurasiatic steppe. None of their languages succeeded in making any sort of permanent change in any great part of Europe. Always the Indo-European languages of the larger settled farming populations win out in the end.

Indeed, it doesn't really matter one bit if one accepts elite dominance or folk migration or demic diffusion as more important in the spread of language in Europe. What does matter is that Europe has been resilient to linguistic change from outside: neither the Arabs, nor the Turks, nor the Mongols, nor the various Iranians, nor the various Uralic speakers that came to it were able to effect real linguistic change.

pconroy said...

Dienekes,

One of the problems with your Anatolian theory is that it fails to account for the spread of IE in Asia.

How did Anatolian languages spread to Sri Lanka and who spread them, and when, if not IE speakers.

Also, how do you account for reconstructed PIE containing words associated with Steppe life?

Ardagastus said...

This is an example of a folk migration, NOT of elite dominance, as there is no evidence that the Celts in either the Balkans or in Anatolia imposed their language on the natives at all, and indeed their language was lost very quickly by that of the larger surrounding population.

Celts arrived in Anatolia in 3rd century BCE as warriors and mercenaries (invited by Nicomedes I in his struggle for the Bithynian throne) and their language survived at least until the 4th century CE (when it is attested by Jerome), which would account for some seven centuries. Also it's impossible that from Britain to Anatolia Celts spread through "folk migration", as there was no endless source of Celts somewhere in Central Europe! Moreover, the same Celtic tribes are attested in Galatia and Western Europe (for instance Tectosagi), showing that only smaller groups migrated, not the entire community, and moreover the contemporary sources attest mostly Celtic armies and warriors, battles and plunders, which is further proof that those migrating Celts were military elites.

I must repeat (because so do you), there's no archaeological and anthropological evidence for massive migrations. Consequently invoking the "folk migrations" is a speculation.

Which makes it even more surprising that absolutely none of the language shifts within Europe (with the exception of Hungarian) originated in the steppe region, the alleged homeland of the PIE.

Not true, large areas of Eastern Romania, Moldavia, Ukraine, Russia became Turkic speaking (erasing the older Iranian, Gothic, Slavic and whatever other languages) for most of the Middle Ages and early modern era until their languages were mostly assimilated by Romanian, Ukrainian and Russian.
However Celtic, Germanic, Latin, Slavic speakers, who caused the indo-Europeanization of most of the Europe, were no longer "steppe warriors" at the time of their expansion.

They did. We have repeated movements of Iranian, Turkic, Mongolian, and Uralic speakers into Europe from the Eurasiatic steppe. None of their languages succeeded in making any sort of permanent change in any great part of Europe. Always the Indo-European languages of the larger settled farming populations win out in the end.

"Permanent" is in the eye of the beholder and most steppe movements you mention affected only regions of Eastern Europe. Tatar is spoken even today in some countries of Eastern Europe (and Mongol invasions brought the Tatars, the Mongol armies retreated).

Most languages are brought by nomads, be them colonists (in the case of Roman Empire) or wandering groups (like the Anglo-Saxon invasion of England).

Indeed, it doesn't really matter one bit if one accepts elite dominance or folk migration or demic diffusion as more important in the spread of language in Europe. What does matter is that Europe has been resilient to linguistic change from outside: neither the Arabs, nor the Turks, nor the Mongols, nor the various Iranians, nor the various Uralic speakers that came to it were able to effect real linguistic change.

Arabs, Turks, Iranians brought linguistic change in Europe. Obviously Europe is not resilient to linguistic change, otherwise we would have spoken the same languages since Paleolithic. Only PCT and similar fringe positions make such arguments.

Dienekes said...

Celts arrived in Anatolia in 3rd century BCE as warriors and mercenaries

It doesn't matter one iota what their profession was. Elite dominance is the process by which individuals assume positions of power through which they are able to spread their language and culture. There is absolutely zero evidence that Celts had a position of power in the Roman Empire in the regions where they were settled. The Galatians were troublesome barbarians that the Romans settled in Anatolia, like the common imperial policy. They practiced regular economic activities in the area, not warfare -as there were no Roman enemies in Central Anatolia.

And, as we all know their language was lost. The fact that it lasted 7 centuries is irrelevant. Obviously whenever a new language is introduced, it will be spoken for some -shorter or longer- time. The point is that none of the languages that were introduced to Europe from the east -and Celtic, BTW, is not such a language- had a lasting influence even in a small region of Europe, let alone the entire continent.

I must repeat (because so do you), there's no archaeological and anthropological evidence for massive migrations. Consequently invoking the "folk migrations" is a speculation.

On the contrary, there is ample anthropological, genetic, and archaeological evidence for migrations of Neolithic farmers from Anatolia to Europe. And no evidence of similar strength regarding the impact of Kurgan people on Europe.

Not true, large areas of Eastern Romania, Moldavia, Ukraine, Russia became Turkic speaking (erasing the older Iranian, Gothic, Slavic and whatever other languages) for most of the Middle Ages and early modern era until their languages were mostly assimilated by Romanian, Ukrainian and Russian.

That is incorrect. Turks were never a majority in any region of eastern Europe. They certainly did not affect at all most of Europe. And, their linguistic presence was short-lived, with the Indo-European elements winning out in the end.

"Permanent" is in the eye of the beholder and most steppe movements you mention affected only regions of Eastern Europe.

Exactly. Historical steppe movements only affect parts of Eastern Europe, and eventually (except in the case of Hungary) fail, and the IE element reasserts itself.

Tatar is spoken even today in some countries of Eastern Europe

Irrelevant. The point is that Turks (and all the other eastern steppe people) did not affect large areas of Europe. That there are small remnants of them is testament to their inability to win out against native IE languages of Europe.

Most languages are brought by nomads, be them colonists (in the case of Roman Empire) or wandering groups (like the Anglo-Saxon invasion of England).

Romans were not "nomads" and neither were Anglo-Saxons. Do you even know what a nomad is?

Arabs, Turks, Iranians brought linguistic change in Europe.

Short-lived change in the periphery of Europe that disappeared within a few centuries. Even though (at least Arabs and Turks) had real conquering armies and a central government co-ordinating their expansion.

Obviously Europe is not resilient to linguistic change, otherwise we would have spoken the same languages since Paleolithic.

That is a red herring. I spoke specifically about linguistic change effected by steppe pastoralists from the east. Clearly there has been linguistic change within Europe (e.g., the spread of Latin or Germanic), and the influx of a sharply genetically differentiated population in Europe during the Neolithic would almost have certainly have brought a new language (Indo-European) into Europe.

One of the problems with your Anatolian theory is that it fails to account for the spread of IE in Asia.

The Anatolian theory is equivalent to the Kurgan theory in terms of the spread of IE in Asia. Renfrew originally proposed two models for the Asian spread: with either early farmers, or the traditional Kurgan theory.

Ardagastus said...

The point is that none of the languages that were introduced to Europe from the east
False, the non-IE Hungarian and Tatar are still spoken today. IE languages were also introduced from east (whether from Anatolia or north-Pontic area).

Celtic, BTW, is not such a language

But there are Celtic languages (even Renfrew, Gray and Atkinson mention them).

On the contrary, there is ample anthropological, genetic, and archaeological evidence for migrations of Neolithic farmers from Anatolia to Europe. And no evidence of similar strength regarding the impact of Kurgan people on Europe.

There's no evidence whatsoever for massive migrations of Germanic, Celtic, Slavic speakers, there's no evidence for the massive prehistoric migrations of IE speakers, consequently when addressing a Germanic, Celtic or Slavic "folk migration" is a wild speculation rooted in wishful thinking.

That is incorrect. Turks were never a majority in any region of eastern Europe. They certainly did not affect at all most of Europe. And, their linguistic presence was short-lived, with the Indo-European elements winning out in the end.

Pechenegs, Uzes, Cumans, Tatars were the dominant population and Turkic was the dominant languages over vast areas of Eastern Europe for centuries, Their language is often the only language attested by the contemporary sources (see Constantine Porphyrogenitus' ethnographic description of Patzinakia) and it even left considerable traces in toponymy. The IE element which won was actually the assimilation policies of the modern states not some vigour of the IE languages spoken by farmers, thus this argument is of no relevance.

Exactly. Historical steppe movements only affect parts of Eastern Europe, and eventually (except in the case of Hungary) fail, and the IE element reasserts itself.

Since steppe movements affected considerably the languages spoken in the territories where these people migrated (your examples were about Eastern Europe), it strongly suggests that this could be also the case for similar prehistoric movements.

Romans were not "nomads" and neither were Anglo-Saxons. Do you even know what a nomad is?


I actually said Roman colonists and of course they were. Do you know what a nomad is?

That is a red herring. I spoke specifically about linguistic change effected by steppe pastoralists from the east.


There's no red herring as you actually said "What does matter is that Europe has been resilient to linguistic change from outside" (nothing about pastoralists) which is a preposterous claim.

Renfrew originally proposed two models for the Asian spread: with either early farmers, or the traditional Kurgan theory.

So the Kurgan theory actually
works when Renfrew uses it. How convenient!

Ardagastus said...

It doesn't matter one iota what their profession was. Elite dominance is the process by which individuals assume positions of power through which they are able to spread their language and culture. There is absolutely zero evidence that Celts had a position of power in the Roman Empire in the regions where they were settled. The Galatians were troublesome barbarians that the Romans settled in Anatolia, like the common imperial policy. They practiced regular economic activities in the area, not warfare -as there were no Roman enemies in Central Anatolia.

Ignorance is no argument. There is plenty of evidence the Galatians arrived in Anatolia in 3rd century BCE as mercenaries (not as traders) invited by Bythinians (no Romans settled them, moreover as there was no Roman Empire in 3rd century BCE there couldn't be any imperial policy!), they had their kings, they participated in many local armed conflicts as mercenaries and allies and they spoke a Celtic language (as contemporary testimonies point out) so all evidence suggests a local Celtic language maintained through elite dominance.

Also Celts in general had positions of power in the Roman Empire through civitas peregrinae and similar institutions. During centuries of Roman domination many of them eventually switched to Latin but not all of them (Roman Britain is a notable example)

The fact that it lasted 7 centuries is irrelevant.

That it lasted seven centuries it is very relevant to show a permanent language shift. When Galatians were assimilated they didn't revert to some ancient Anatolian language but they learnt Koine Greek (and later that area became Turkic speaking as it is today).

Maju said...

The only worthwhile examples are of Arabs and Turks who spread because of their organization, i.e., they were organized states (Khalifate, various Turkish sultanates) which pursued a policy of conversion over many generations.

In fact the Ummayad Caliphate did not pursue conversion (specially as that meant that subjects would not pay taxes anymore). The Ottoman Empire didn't either, at least not too much, (as they relied heavily on Christian populations to feed their janissary infantry). Just that certain sectors, (Arabized or Turkicised) Muslims in these cases, were clearly privileged and there was a gate open to join their ranks. But that happened in all Muslim polities and, equally in all polities that prime one identity over other others.

But Turks were expanding and assimilating peoples long before they became Muslims. And Akkadians, Persians, etc. never had a religious discrimination policy that I know of.

Even the Iroquois, a quite interesting historical example of matrifocal neolithic society, when defeated some enemy offered them the opportunity to join their league (or emigrate). Joining the Iroquois would of course suppose renouncing partly to their original identity and become assimilated, even if as a tribe with equal rights.

But the age of Indoeuropeans (Chalcolithic and further on) was more advanced in most aspects and was evolving from mere tribalism into the proto-feudalism that surely existed in the Iron Age. People think of La Tène Celts as "tribes" but apart of having some tribal structure, they were also organized as city states (oppidae per the Roman accounts) ruled by monarchs and with strong social inequality: slavery and serfdom, as well as aristocracy, existed already and were not just anecdotal.

And that was already being formed in the Chalcolithic.

... no evidence of a great Kurgan "state" whatsoever.

Ok, so the Baalberge culture, that opened and colonized the, till then uninhabited, huge Brandenburg forest, conquered Cuyavia and northern Moravia and had a capital town with clear signs of "rural" aristocracy, were just a band of warriors with no organization whatsoever.

I'm pretty sure they were tribal and bellicose but also that they had some sort of political organization as well, maybe an elected monarchy like the kind we see in so many historical "tribal" aristocratic societies. That is enough, as you can see in Anglosaxon England for instance.

Even Slavs that were initially pretty reluctant to high-level political organization, managed to assimilate wide areas without such tool. Hungarians? Just feared tribal hordes that raided half Europe... Well, they happened to have a monarchy.

Each case is surely unique in many details but elite dominance and political organization can perfectly happen (and in fact happened almost all the time) in Metal Ages "tribal" societies.

Maju said...

Neither Brennus nor Attila made even a tiny change in a small part of the linguistic landscape of Europe, so I don't see why much more primitive Kurgan people would have changed the languages of the whole of Europe.

They are examples of the political organization that "tribal" societies can achieve. Certainly Brennus consolidated the Celtization of Northern Italy that "other Brennus" before them had estabilished just decades or centuries earlier.

How come the people of Roman Modena could claim being of Etruscan origins if all the Emilia had been conquered by the Gauls long before? If, as you seem to imagine, each conquest implied a total ethnic cleansing and new settlement, there was no way a single Etruscan remained in the city. History tells us otherwise. Would enough time have passed, they would even had forgotten their Etruscan ancestral identity though.

Latin spread because it was the language of an organized state that pursued a policy of expansion for many centuries.

In fact Latin (including vulgar Latin) consolidated its position, it seems, in the Dark Ages, after Rome was gone. It was the language of the elites (even if these elites were often Germanic they assumed Latin as the state language) and that kept Latin expanding even after Rome was dead. Only in the rare places where no state that claimed itself the local heir of Rome (Goths and Franks legitimized their rule on grounds of vassalage to Rome and both used the Catholic Church as tool of their power) existed, like Vasconia, Britain or the Slavic territories of the Balcans, Latin lost positions or took longer to consolidate.

Think North Africa: it was as Western Roman as anywhere else but Berber, not Latin, persists from that time. The Western Empire was not fully Latinized when Rome fell. The process of Latinization was completed after Rome succumbed - under "tribal" states (Franks and Goths deserve no other name).

Well... we could go for centuries like this: c. 200 posts is enough for me. It's a fascinating debate but I doubt we can persuade Dienekes (nor vice versa) and this format is kind of limiting too... Guess I'm done with this discussion.

terryt said...

"The fact that it lasted 7 centuries is irrelevant".

Would we call it 'irrelevant' if we were having this discussion during those 700 years?

"And, their linguistic presence was short-lived, with the Indo-European elements winning out in the end".

But often not the IE language they had originally replaced. In many cases a new one introduced from outside the particular region

"the influx of a sharply genetically differentiated population in Europe during the Neolithic would almost have certainly have brought a new language (Indo-European) into Europe".

Would we be able to discern any relationship at all within what we know as the Indo-European language family if its spread was really that ancient?

Dienekes said...

False, the non-IE Hungarian and Tatar are still spoken today. IE languages were also introduced from east (whether from Anatolia or north-Pontic area).

Are you deliberately obtuse? Hungarian and Tatar did not affect most of Europe, and I specifically spoke about the impact of steppe elements from the east. We have absolutely no data points about migrations from Anatolia, but we have several data points about migrations of steppe elements from the east, none of which effected the change postulated for "Kurgan" folk.

But there are Celtic languages (even Renfrew, Gray and Atkinson mention them).

Celtic languages did not originate in the steppe. Celts were not steppe pastoralists. Nor did they impose their language on Anatolia by elite dominance. Your example is muddle-headed on at least three counts. The example adds zero evidence in favor of the idea that Kurgan folk could have spread their language into Europe.

There's no evidence whatsoever for massive migrations of Germanic, Celtic, Slavic speakers

Irrelevant, since all these were within-Europe population movements of farmers/stockbreeders, not intrusions into Europe by steppe warriors as the postulated "Kurgan" expansions.

Pechenegs, Uzes, Cumans, Tatars were the dominant population and Turkic was the dominant languages over vast areas of Eastern Europe for centuries

Nonsense, these were small tribal units that attracted attention because they were troublesome to the Roman Empire. Most Eastern Europeans spoke Indo-European languages (principally Slavic ones) and the various intruders disappeared without a trace.

I actually said Roman colonists and of course they were. Do you know what a nomad is?

I sure as hell know what a nomad is, and Roman colonists were not nomads.

There's no red herring as you actually said "What does matter is that Europe has been resilient to linguistic change from outside" (nothing about pastoralists) which is a preposterous claim.

It is not a preposterous claim at all as neither today nor at any time in its know history has any great part of Europe spoken a language that originated outside the continent. This is not true for the Eurasiatic steppe, or for the Levant, or for North Africa, but it is true for Europe.

Dienekes said...

So the Kurgan theory actually
works when Renfrew uses it. How convenient!


Renfrew uses it for the eastern expansion of Indo-Iranians (and perhaps Tocharians), which is the only fairly documented one, given the close relationship between Indo-Aryans and Iranians, and the historical presence of Iranian speaking Scytho-Sarmatian type people in the steppe.

He does not, of course use it for Europe and Anatolia where there is absolutely no evidence for linguistic change effected from the steppe

Ignorance is no argument. There is plenty of evidence the Galatians arrived in Anatolia in 3rd century BCE as mercenaries (not as traders) invited by Bythinians (no Romans settled them, moreover as there was no Roman Empire in 3rd century BCE there couldn't be any imperial policy!), they had their kings, they participated in many local armed conflicts as mercenaries and allies and they spoke a Celtic language (as contemporary testimonies point out) so all evidence suggests a local Celtic language maintained through elite dominance.

Stupidity is no argument either. Barbarian warrior peoples were often invited into Roman territory, but being warriors is only one part of the equation of "elite dominance". The other part is "dominance". The Galatians did not dominate Roman society despite being "warriors".

Also Celts in general had positions of power in the Roman Empire through civitas peregrinae and similar institutions.

Nonsense, Celts had no special position in the Roman Empire. Certainly no evidence at all that they "dominated" either the Roman Empire itself or any part thereof. After they -like other subject peoples- were granted Roman citizenship, they could hold various offices, but there is no evidence that they were in anyway dominant, and in any case it was the ones that had knowledge of Latin (or Greek in the east) that could really hold such position. It is laughable to think that monolingual Celts spread their Celtic language by being powerful officers of the Roman Empire!

In fact the Ummayad Caliphate did not pursue conversion (specially as that meant that subjects would not pay taxes anymore). The Ottoman Empire didn't either, at least not too much, (as they relied heavily on Christian populations to feed their janissary infantry).

Both states' policies facilitated conversion, since non-Muslims had every incentive of abandoning their relgion in order to gain posts and avoid the head tax. It is no accident that Christians have all but disappeared from the vast majority of Muslim states, and with them their languages (Syriac, Coptic, Greek, Assyrian, Armenian, etc.)

But Turks were expanding and assimilating peoples long before they became Muslims.

But not in Europe or Anatolia.

I'm pretty sure they were tribal and bellicose but also that they had some sort of political organization as well

I don't share your certainty as there is nothing to support a "Kurgan state", and the fact that the Indo-Europeans -whatever their origin- splintered off into a dozen different language groups, each of them fighting each other, suggests rather that there was no central powerful authority (like the Sultan or Khalif) co-ordinating their expansion.

In fact Latin (including vulgar Latin) consolidated its position, it seems, in the Dark Ages, after Rome was gone. It was the language of the elites (even if these elites were often Germanic they assumed Latin as the state language) and that kept Latin expanding even after Rome was dead.

Latin did not expand after Rome was dead. The limits of Latinity are pretty much the limits of the Roman Empire, and in the periphery Latin contracted (e.g., in Britain, North Africa, Central Europe, and the Balkans).

Dienekes said...

Would we call it 'irrelevant' if we were having this discussion during those 700 years?

There is no known span of 700 years in which any great part of Europe spoke anything other than Indo-European languages.

But often not the IE language they had originally replaced. In many cases a new one introduced from outside the particular region

But none outside Europe, except for the eastern edge of Europe, in the border of the Eurasiatic steppe where Iranian-type languages gave way to Slavic-type ones in historical time.

Would we be able to discern any relationship at all within what we know as the Indo-European language family if its spread was really that ancient?

Why not? Semitic is Bronze Age, and Afroasiatic is believed to be Neolithic or even earlier, but linguists have no trouble discerning the relationships between them. Moreover, if we exclude Hittite, the rest of the IE languages have more recent common ancestry than the initial Neolithic colonization of Europe.

«Oldest ‹Older   1 – 200 of 204   Newer› Newest»