April 08, 2011

Indo-European origins: Neolithic Anatolia still the best hypothesis

The recognition that Bayesian phylogenetic methods, first developed in biology, could also be applied to the evolution of languages, has been one of the most exciting developments over the last decade.

This realization, is due in a great part to Russell Gray and Quentin Atkinson (henceforth G&A) and their influent 2003 Nature article, in which they argued that a robust estimate, consistent with the Anatolian hypothesis of Indo-European origins (left), could be inferred from lexicaldata.

The G&A was initially met with skepticism because it was conflated with glottochronology, a now controversial method of inferring language time depth based on rates of retention of cognates. Glottochronology tried to apply a regular law to linguistic change, a notoriously irregular process. As the authors of the current paper note:
glottochronological calculations with considerable
scepticism. The most fundamental obstacle
encountered by glottochronology is the fact that
languages, just like genes, often do not evolve at a constant
rate.
In their classic critique of glottochronology, Bergsland & Vogt [12] compared present-day languages with their archaic forms. They found considerable evidence of rate variation between languages. For example, Icelandic and Norwegian were compared with their common ancestor, Old Norse, spoken roughly 1000 years ago. Norwegian has retained 81 per cent of the vocabulary of Old Norse, correctly suggesting an age of approximately 1000 years. However, Icelandic has retained over 95 per cent of the Old Norse vocabulary, falsely suggesting that Icelandic split from Old Norse less than 200 years ago.

The critics largely ignored the fact that G&A's method avoided the problems of glottochronology, such as the questionable assumption of a constant rate of evolutionary change: instead, the G&A method exploited multiple known calibration points (e.g., the breakup of Romance languages in Late Antiquity) and did not need such a strict and unrealistic assumption.

The authors of the current paper bemoan the strange reversal of fortunes of computational methods in biology and linguistics. Whereas the attempt to estimate dates from variation has a longer pedigree in linguistics than in biology (and I would add that it was even older than the radiocarbon revolution in archaeology), computational methods have won the day in biology, and been abandoned by linguists:
It is ironic that over the past half-century, computational methods in historical linguistics have fallen out of favour while in evolutionary biology computational
methods have blossomed. Rather than giving up and saying, ‘we don’t do dates’, computational biologists have developed methods that can accurately estimate phylogenetic trees and divergence dates even when there is considerable lineage-specific rate heterogeneity.

Much of the criticism against G&A stemmed from the received "knowledge" that Indo-European could not have been as old as the Neolithic spread from Anatolia to Europe. And, yet, if you ask linguists why they are so sure that this is the case, you will mostly get either (i) a non-quantitative opinion (i.e., an ex cathedra guess, since linguists "don't do dates"), or (ii) a redirect to archaeology. Ask archaeologists who ascribe to the same opinion, and you will either get a redirect to "linguists", completing the loop, or an inexplicable belief in the ability of Chalcolithic pastoralists from the Eurasian steppes of effecting almost total linguistic replacement over a huge area from the Atlantic to India, despite:
  • meagre evidence for their actual presence outside the steppes and a few bordering areas, such as the Danubian region, and
  • a total lack of knowledge of what languages early steppe pastoralists spoke, assuming that because Scythians spoke (most likely) an Iranic language 2-3 thousand years later, so must have the early steppe populations

What followed the publication of the 2003 article was a remarkable back-and-forth between G&A and their critics, in which one by one the technical objections to their discovery was addressed, and computational linguistic techniques reached maturity by being applied in numerous different ways and on different datasets for the problem of not only Indo-European, and on different language families, such as Semitic, Austronesian, Melanesian, the languages of the Sahul, and Arawak.


The new paper is a nice summary of what has transpired over the last eight years. G&A together with Simon Greenhill detail how one by one the skeptics' arguments have been addressed:
  • Both the Dyen et al. dataset mostly of modern Indo-European languages as well as an independent one by Ringe et al. on mostly ancient ones gave similar results
  • Binary vs. multi-state coding of cognate sets gave similar results
  • By removing calibration points they estimated the age of nodes of known age, coming up with results close to the truth and, in any case, not systematically older
  • An application of Dollo's Law on both the Ringe et al. and Dyen et al. data, i.e., higher rate of cognate gain than cognate loss did not affect the time depth of the estimate
  • Removing some calibration points, or cognates dubbed as dubious in the data, or of limiting the analysis to the most stable ones didn't affect the results, and sometimes made them appear older
I've often said that linguistics is dangerous territory, as a handful of people are probably proficient in all the languages involved, which includes not only several IE sub-families, but also Semitic, Finno-Ugrian, and NE/S Caucasian. Nonetheless, I can't help but notice that G&A and their colleagues have done an exceptional job of addressing objections to their work in a replicable and quantitative manner.

Archaeological and Genetic evidence

The Anatolian hypothesis has been challenged on non-linguistic as well as linguistic grounds. In terms of archaeology, there is a school of thought that favored static models of cultural change and was allergic to migration. However, recent work on both the craniometry and the DNA of ancient farming communities in Europe in comparison to the Mesolithic population is supportive of a population influx.

Surprisingly, this influx does not seem to be a good match to either static models or the classical demic diffusion hypothesis, in which farmers advance in small steps, gradually intermingling with foragers before moving on further away from their region of origin. Instead, early farming communities in LBK largely avoided the foragers and vice versa, an idea that is borne out by the surprisingly large measurable genetic differences between the two populations.

Hence, there is clear evidence of a population transfer during the onset of the Neolithic; once fashionable ideas like: "Mesolithic people received grains and pots, but not people and languages from West Asia" are not valid. We cannot be sure what languages the European proto-farmers spoke, but, their path is consistent with the Anatolian hypothesis, and provides a parsimonious explanation for the oldest split between Anatolian languages and Indo-European.

Conclusion

The Anatolian hypothesis of Indo-European origins is very much alive, and, in my opinion, still the best explanation for the fact that people from the Atlantic to the Indian Ocean and the outskirts of China speak/spoke languages stemming from a common ancestor.

Related:
  • Reali and Griffiths recently made a strong argument about the applicability of biological methods to language evolution, considering the different (genetic vs. "iterated learning") modes of transmission
  • A common ancestor of Indo-European and Hurrian?
  • Lexican borrowing in the history of Indo-European languages
  • The shape and tempo of language evolution

Phil. Trans. R. Soc. B 12 April 2011 vol. 366 no. 1567 1090-1100

Language evolution and human history: what a difference a date makes

Russell D. Gray, Quentin D. Atkinson and Simon J. Greenhill

Abstract

Historical inference is at its most powerful when independent lines of evidence can be integrated into a coherent account. Dating linguistic and cultural lineages can potentially play a vital role in the integration of evidence from linguistics, anthropology, archaeology and genetics. Unfortunately, although the comparative method in historical linguistics can provide a relative chronology, it cannot provide absolute date estimates and an alternative approach, called glottochronology, is fundamentally flawed. In this paper we outline how computational phylogenetic methods can reliably estimate language divergence dates and thus help resolve long-standing debates about human prehistory ranging from the origin of the Indo-European language family to the peopling of the Pacific.

Link

87 comments:

Måns Sjöberg said...

One problem with the Anatolian hypothesis is that is fails to explain how come words for milk, wheel, wool, and ard were part of the Protoindoeuropean lexicon. These "secondary neolithic products" were invented much later.

German Dziebel said...

"We cannot be sure what languages the European proto-farmers spoke, but, their path is consistent with the Anatolian hypothesis, and provides a parsimonious explanation for the oldest split between Anatolian languages and Indo-European."

People tend to focus on the westward spread of IE languages, as if it's the only thing that needs to be explained. How about the sudden emergence of Tocharian all the way in the east, Tocharian being the second earliest branch of IE after Anatolian? Did a bunch of early eastward-expanding IE languages go extinct and the link between Anatolia and China got severed? Also, Indo-Iranian languages are fairly close to the ancestral node, and they are also eastward from Anatolia and could provide the missing link between Anatolian and Tocharian. But then one would expect an earlier branch to be closer to the source geographically, but Tocharians are further removed from Anatolia than Indo-Iranians. In the westward direction, the phylogeny seems to replicate geographic proximity, with Greek and Armenian branching off before Celtic just like the Balkans are closer to Anatolia than the British Isles.

Albanians cluster with Indo-Iranians as if population influxes going through the Balkans and in the eastern direction (Tocharian <-> Greek, Indian <-> Albanian) were correlated at least twice in the history of IE. Possible of course but somewhat suspicious. Also, I'm not aware of any isoglosses connecting Albanian exclusively to Indo-Iranian.

Another thing to consider: if Hittite and Tocharian weren't discovered (and their discovery was a historical accident), this cool methodology would furnish a "proof" that Greeks came from Armenia, Indians from Albania, etc. History is a sequence of unique events and quantitative methods are only as successful as the underlying facts, which they purport to explain, allow them to be. So, when the authors and Dienekes accuses linguists of being non-receptive to the new computer games they put the cart before the horse. It's quantitative methods that need to catch up with the complexity of cultural data spread out in time and space, not the other way around.

Dienekes said...

The secondary products revolution is largely a failed hypothesis. The dates for milk use are consistent with the dates of the G&A model

http://dienekes.blogspot.com/2008/08/earliest-evidence-for-milk-in-near-east.html

As for "wheel" it is a topic that has been addressed to death. No need to infer the presence of "wheels" in PIE, just of a common word for "round" or "twist", and nature provides many examples of round things. Indeed, wheels themselves could've been used much earlier than their use for transport (as toys, for example).

Dienekes said...

How about the sudden emergence of Tocharian all the way in the east, Tocharian being the second earliest branch of IE after Anatolian? Did a bunch of early eastward-expanding IE languages go extinct and the link between Anatolia and China got severed?

I am not sure why you call Tocharian early. It is only attested in the first millennium AD. If you mean that it _split_ early, this is probably true, although the path it took to arrive in China is hard to detect.

Certainly I don't see any difficulty with a pre-Indo-Iranian eastern spread of some Indo-European group from either tha Balkans or directly from West Asia. I always compare Tocharians with Magyars: a lately-attested interesting isolate surrounded by people of different language.

It could also be the case that Tocharian may be a remnant of _early_ non-Indo-Iranian eastern Indo-Europeans. Iranian was largely replaced by Turkic over much of Asia in a thousand years, so it's quite possible that Iranian had previously replaced Tocharian-like languages in Asia.

Another thing to consider: if Hittite and Tocharian weren't discovered (and their discovery was a historical accident), this cool methodology would furnish a "proof" that Greeks came from Armenia, Indians from Albania, etc.

The Anatolian split already puts PIE at the limits of what can be inferred.

German Dziebel said...

"The Anatolian split already puts PIE at the limits of what can be inferred."

My point is it doesn't seem correct to lump together a finite set (currently attested IE languages) and an infinite set (accidentally discovered ancient languages such as Hittite and Tocharian, and there could be any number of those). And then make a prediction on the basis of the infinite set. At the level most divergent ancient, extinct languages, Tocharian and Hittite would form a binary split with no way to decide whether the Hittites came from China or Tocharians cam from Anatolia.

"The limits of what can be inferred" doesn't apply to ancient, extinct languages. There can be any number of those in any place along the continuum from China to Anatolia.

If we exclude Tocharian and Hittite and use only existing languages, then what we can infer is that the primary branches are Greek, Armenian, Albanian and Indo-Iranian, which again span a vast geographic terrain, with several possible points of origin.

Once again, it's a rather spurious exercise.

I do find it interesting that there are 3 layers of east-west linguistic ranges within IE: Slavic <-> Celtic, Indo-Iranian <-> Albanian-Greek and, on the level of ancient extinct languages Tocharian <-> Hittite. But the level of ancient extinct languages is not robust enough to infer the ultimate origin of this IE "funnel."

Dienekes said...

My point is it doesn't seem correct to lump together a finite set (currently attested IE languages) and an infinite set (accidentally discovered ancient languages such as Hittite and Tocharian, and there could be any number of those).

I see no reason why not.
There are no "infinite sets" by the way, both sets are finite.

Going by your "logic" evolutionary biologists would not lump together extinct and extant taxa to try to discover the evolutionary history of all life, since the fossil record may not contain all life forms that ever lived on the planet.

At the level most divergent ancient, extinct languages, Tocharian and Hittite would form a binary split with no way to decide whether the Hittites came from China or Tocharians cam from Anatolia.

Hittites did not come from China, because of parsimony. I hope you won't start arguing that IE languages originated in East Asia; it's bad enough that you think that mankind originated in the Americas...

German Dziebel said...

"There are no "infinite sets" by the way, both sets are finite."

One set is finite, the other one is infinite. There are no more IE languages but the ones currently spoken, but there's an infinitely possible number of languages such as Hittite and Tocharian that went extinct. One more find of an ancient IE language in China and the whole model collapses. The whole model hinges on an extinct language (or a group of closely related languages) but the fact that it got extinct makes other extinctions possible, hence any of them could falsify the primacy of the Anatolian branch. It's a circular argument. How can you argue with the obvious?

"Going by your "logic" evolutionary biologists would not lump together extinct and extant taxa to try to discover the evolutionary history of all life, since the fossil record may not contain all life forms that ever lived on the planet."

If biologists had two extinct, widely-dispersed, disjointed and poorly preserved taxa and a dozen of extant and densely populated taxa, they would be very careful to declare where the ancestors of all taxa lived.

"Hittites did not come from China, because of parsimony."

Your primitive parsimony doesn't apply to a historical process. They could come from anywhere.

"I hope you won't start arguing that IE languages originated in East Asia"

Dienekes, educate yourself before jabbing. Johanna Nichols's model, which is not an ex cathedra opinion, places IE homeland in Bactria, which is pretty close to China. See here. http://books.google.com/books?id=s3a-M8btiQoC&pg=PA485&lpg=PA485&dq=johanna+nichols+bactrian+IE&source=bl&ots=0NaPTLU14_&sig=Edrc2j72mQSihYd6GTOkf2czRmM&hl=en&ei=w4yfTY2bDqLz0gGHz-iiBQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CBQQ6AEwAA#v=onepage&q=johanna%20nichols%20bactrian%20IE&f=false

"it's bad enough that you think that mankind originated in the Americas."

Out of America explains the observed genetic, kinship and linguistic variation better than out of Africa and it fits archaeology like hand and glove. At this point I don't even know what YOUR theory of human origins is. Humans came from everywhere and from nowhere? How is better than out of America?

Dienekes said...

One set is finite, the other one is infinite.

Do you know what infinite means? Even if every human that ever lived on the planet spoke their own language, it'd still be a finite set.

One more find of an ancient IE language in China and the whole model collapses. The whole model hinges on an extinct language (or a group of closely related languages) but the fact that it got extinct makes other extinctions possible, hence any of them could falsify the primacy of the Anatolian branch. It's a circular argument. How can you argue with the obvious?

Every scientific theory is falsifiable. The fact that a new Indo-European language in China can falsify the model does not mean that it's not the best model there is.


Your primitive parsimony doesn't apply to a historical process. They could come from anywhere.

Parsimony applies to all scientific theories.


Dienekes, educate yourself before jabbing. Johanna Nichols's model, which is not an ex cathedra opinion, places IE homeland in Bactria, which is pretty close to China.

Bactria isn't part of East Asia by any reckoning. Having read the theory, I see absolutely no value to it. You're entitled to believe in it, however, it's marginally more believable than your Chinese hypothesis.

Out of America explains the observed genetic, kinship and linguistic variation better than out of Africa and it fits archaeology like hand and glove. At this point I don't even know what YOUR theory of human origins is. Humans came from everywhere and from nowhere? How is better than out of America?

Out of America is a nonsensical theory with zero evidence in its support. It's actually the second worst theory of human origins, with the absolute worst being that they came from outer space...

German Dziebel said...

"Do you know what infinite means?"

Yes. But you don't. Let me explain it to you: a finite set is a set that has a finite number of elements. [1, 2, 3] is a finite set. The opposite is an infinite set, It has an infinite number of elements [1,2,3...]. Easy enough?

"The fact that a new Indo-European language in China can falsify the model does not mean that it's not the best model there is."

No model can be based on a single extinct taxon. Tocharian had already pre-falsified the Anatolian model before it was even advanced.

"Parsimony applies to all scientific theories."

You're just mudslinging words the meaning of which escapes you.

"Bactria isn't part of East Asia by any reckoning. Having read the theory, I see absolutely no value to it. You're entitled to believe in it, however, it's marginally more believable than your Chinese hypothesis."

I don't have any Chinese hypotheses. I'm just monitoring the data as it comes in and weeding out false prophets. Under Nichols's Baktrian hypothesis, Tocharian is the language that stayed behind. So it's opposite interpretation of the basal structure of the IE phylogeny from yours and some other Anatolianists.

"I see absolutely no value to it."

Dienekes, with all due respect, you can supply the data and fresh insights. But professionals will handle the interpretations.

"Out of America is a nonsensical theory with zero evidence in its support. It's actually the second worst theory of human origins, with the absolute worst being that they came from outer space..."

You are just ignorant. It will take you years to re-educate yourself, so I don't expect you to. Plus I like you the way you are.

Jesús Sanchis said...

I agree that the Anatolian hypotyhesis is based on solid grounds and must be taken into account in any study of IE origins. However, Gray and Atkinson's approach is useless. It is based on a series of assumptions that can easily be challenged. That's the problem with computerized calculations of human affairs: the computer seems to put things in order and offer a clear account of the data from an objective perspective, but the people who feed the data are always biased. A firm belief in language genealogical trees is an example of bias.

As for words like 'wool', it is obvious that some linguists have tried to explain its etimology in terms of 'wool that is obtained from domesticated animals', but that could be just wishful thinking. Wool appears in nature. For any other supposed common PIE agriculture-pastoralist language, an alternative explanation can be offered, including diffusion.

eurologist said...

I will read the paper later, here just a few comments:

- The old dates of the major splits are refreshing - surely much closer to the truth than what is often promulgated. As I have often said, with dates moving back farther every decade, it becomes harder and harder to ignore the evident coupling of the spread of IE and agriculture.

- The huge temporal gaps between the oldest branches and modern branches surely are some side-effect of the methods used. One concrete example: while Germany, the Netherlands, and Scandinavia obviously form a Sprachbund, that does not mean they only split 1,750 ya. In fact, some of the German dialects alone likely have their origin 500 years before then - as is attested by the huge differences between, e.g., Gothic, Frisian, and Southern Rhineland German during Roman times. The same is reflected in the earliest German literature.

- Similarly, we know that there were already huge differences in Italic languages during early Roman times (500 to 1,000 years before the claimed split at 1,700 ya), and in fact the standard Roman Latin did not prevail in most regions, preserving other, earlier dialects, instead.

- There is still the lingering criticism that Hittite and Tocharian are not know well enough to hinge an entire tree on them.

- Greek and Armenian surely didn't evolve in isolation over 7,300 years. In fact we know Greek forms a Sprachbund with several Balkan languages.

In summary, the results are also compatible with the view that IE originally is a northern Balkan, East Danubian language, or at least one heavily dominated by the multi-cultural environment of that region during the ~500 year hiatus (of the further spread of agriculture north from there) - with perhaps only sparse impact of Anatolian and early proto-Semitic vocabulary (and none of their grammatical constructs).

AdygheChabadi said...

Are we talking about the ultimate origins of the Indo-Europeans???

I am still very much persuaded by Marija Gimbutas' Kurgan Hypothesis...But would that be the origins of the Indo-Iranians or the ultimate origins of Indo-Europeans in general?

Forgive my novice questions...I have never really been as interested in Indo-European origins as I have been other language families like Tyrsenian, Afro-Asiatic, and the non-Indo-European and non-Semitic languages of the ancient greater Middle East.

sykes.1 said...

The comments are too much fun. Everyone knows men come from Mars and women come from Venus.

Anyway, has anyone done for the Anatolian hypothesis what Anthony did for the Kurgan hypothesis in "The Horse, the Wheel, and Language"?

More fun: some people believe that eastern Armenia is the actual site of the Garden of Eden.

Average Joe said...

Does this mean that Indo-European was brought to Europe by R1b males since it is claimed that this particular Y-chromosome haplogroup also came from Anatolia?

Dienekes said...

Yes. But you don't. Let me explain it to you: a finite set is a set that has a finite number of elements. [1, 2, 3] is a finite set. The opposite is an infinite set, It has an infinite number of elements [1,2,3...]. Easy enough?

Yes, and as there is a finite number of atoms on the planet Earth, and a finite number of individuals who belonged to the species Homo sapiens throughout the life of the Earth, by necessity, there has been a finite number of languages spoken by them.

Easy enough?

No model can be based on a single extinct taxon. Tocharian had already pre-falsified the Anatolian model before it was even advanced.

Gee, I wonder if you'd apply your standard of scientific evidence to your Out-of-America theory of human origins. Perhaps there is some extinct population in some island of the Pacific that we haven't sampled yet, so humans really originated in some tropical Pacific island and not in the Americas. And, hey, why stop there, we haven't sampled Mars, perhaps humans originated in Mars....


I don't have any Chinese hypotheses. I'm just monitoring the data as it comes in and weeding out false prophets. Under Nichols's Baktrian hypothesis, Tocharian is the language that stayed behind. So it's opposite interpretation of the basal structure of the IE phylogeny from yours and some other Anatolianists.

Nichols' hypothesis is not parsimonious because it proposes a mysterious drang nach Westen for Proto-Indo-Europeans. Parsimony dictates that Indo-European dispersals would not have a directional preference, and, barring natural obstacles, would radiate in all directions from the homeland.

Plus, there is no archaeological evidence for it. On the contrary, pastoralism spread West->East across Eurasia, and agriculture spread from West Asia. There is absolutely no reasonable agent that could have affected a replacement of languages all the way to the Atlantic from an origin in Bactria.

However, Gray and Atkinson's approach is useless. It is based on a series of assumptions that can easily be challenged. That's the problem with computerized calculations of human affairs: the computer seems to put things in order and offer a clear account of the data from an objective perspective, but the people who feed the data are always biased. A firm belief in language genealogical trees is an example of bias.

Language genealogical trees explain recent language evolution. It's not a belief, it's what has actually happened in a lot of language families in full light of history (e.g., the breakup of Romance languages from Latin).

The only problem with tree models is lateral borrowing. Undetected lateral borrowing would, however, deflate age estimates, as it would make languages appear more closely related than they really are.

Anyway, has anyone done for the Anatolian hypothesis what Anthony did for the Kurgan hypothesis in "The Horse, the Wheel, and Language"?

You mean write a book that largely avoided explaining the spread of most Indo-European languages and focused on a few prehistoric Eurasian cultures assuming that they spoke IE, just because Scythians happened to speak IE in the same area 2 thousand years later?

German Dziebel said...

"Yes, and as there is a finite number of atoms on the planet Earth, and a finite number of individuals who belonged to the species Homo sapiens throughout the life of the Earth, by necessity, there has been a finite number of languages spoken by them.

Easy enough?

No model can be based on a single extinct taxon. Tocharian had already pre-falsified the Anatolian model before it was even advanced.

Gee, I wonder if you'd apply your standard of scientific evidence to your Out-of-America theory of human origins. Perhaps there is some extinct population in some island of the Pacific that we haven't sampled yet, so humans really originated in some tropical Pacific island and not in the Americas. And, hey, why stop there, we haven't sampled Mars, perhaps humans originated in Mars...."

Dienekes, you are just mumbling now. I don't even know what you're talking about. Out of America is a testable hypothesis with just as much cultural and biological support as out of Africa. It's just a matter of how we interpret the data. The data by itself is blind to where humans came from. The fact that the two basal branches of IE - Anatolian and Tocharian - are so widely apart from each other geographically and both are extinct with no living descendants makes it impossible to favor one geographic extreme over the other. We would need another ancient language or a well-reconstructed modern family that would form a transitional stage between IE and the next highest grouping to see if the eastern or the western pole of the IE tree is indicative of the homeland. If, say, IE proves to be closer to Altaic, then it would favor the Bactrian, Chinese or any other eastern homeland. If IE is closer to, sat, Afroasiatic, then it would favor the Anatolian homeland. As of now it's impossible to say. So let's not pretend we can.

"Nichols' hypothesis is not parsimonious because it proposes a mysterious drang nach Westen for Proto-Indo-Europeans."

The older the migration the more likely there won't be any traces of intermediate steps. They will all be wiped out by subsequent migrations, especially in such region as continental Eurasia. Look at the disjointed pattern in the distribution of Y-DNA hg E (SSAfrica) and D (Tibet, Andamans, Japan). Tocharian (barely) survived in a typical refugium zone. Also, if you look at the distribution of Turkic languages, the greatest diversity thereof is in the West but they originated far in the East. The pattern for IE languages may be similar. The Balkans is the most diverse IE region, but this may be due to it being a sink for population movements, not a source.

"Plus, there is no archaeological evidence for it."

Let's stay with linguistics. To quote from you, Dienekes, "And, yet, if you ask linguists why they are so sure that this is the case, you will mostly get either (i) a non-quantitative opinion (i.e., an ex cathedra guess, since linguists "don't do dates"), or (ii) a redirect to archaeology. Ask archaeologists who ascribe to the same opinion, and you will either get a redirect to "linguists", completing the loop..."

Dienekes said...

Does this mean that Indo-European was brought to Europe by R1b males since it is claimed that this particular Y-chromosome haplogroup also came from Anatolia?

I see no real evidence in support of that claim. Also, R1b did not turn up in the (admittedly limited) Central European Neolithic sample of 3 individuals that we have so far. Most importantly, R1b has hardly a presence in India, so if we accept the theorem that Proto-Indo-European patrilineages must've survived even in small percentages in every Indo-European speaking population, India is difficult to explain.

In short, I believe that R1b has played absolutely no role in the early dispersal of the Indo-Europeans.

German Dziebel said...

"On the contrary, pastoralism spread West->East across Eurasia, and agriculture spread from West Asia. There is absolutely no reasonable agent that could have affected a replacement of languages all the way to the Atlantic from an origin in Bactria."

Languages and populations used to spread before agriculture and pastoralism came about. Why do you need a special agent in the form of a food-producing economy to drive it? What agent drove Tocharians from Anatolia to China shortly after the split of Proto-IE? Wind?

"Parsimony dictates that Indo-European dispersals would not have a directional preference, and, barring natural obstacles, would radiate in all directions from the homeland."

This is absurd. In the Neolithic, there were plenty of human barriers to a spread, in addition to natural. Why would a population necessarily radiate in all directions? Humans populations aren't bombs, they don't explode in all directions. Human migrations are driven by economic and sociopolitical factors and those can be skewed one way or another. Under the Anatolian hypothesis you still have to explain the emergence of one of the two most basal branches in China. There's no difference, from the parsimony perspective, which way one of the earliest founding populations went. We can choose a midpoint between the ranges of the two most basal branches, and then derive Tocharian and Anatolian from there. And you'll have your bomb explosion-like dispersal, to boot. But the main thing is that Gray and Atkinson model uses a non-uniform data set composed of a finite set and an infinite set. The Anatolian homeland can't be replicated on the basis of the finite set (There're no extant IE languages in Anatolia). It only transpires on the level on the infinite set. But this infinite set is composed of only two elements - Anatolian and Tocharian, out of possible unattested many. Hence, the homeland inference from Gray and Atkinson's phylogeny is invalid.

"Does this mean that Indo-European was brought to Europe by R1b males since it is claimed that this particular Y-chromosome haplogroup also came from Anatolia?"

What about Klyosov's idea that the spread of IE is associated with Y-DNA R1a (and the Turkic spread with R1b)? This also favors Central/East Asian source for IE languages. http://www.eupedia.com/forum/showthread.php?26108-Thoughts-from-Klyosov-R1b-data-and-the-IE-problem

Jaska said...

Let's play for a while that there is nothing wrong with the dating. Let's just ask: what is the basis for locating PIE in Anatolia? For all that I can see, it's just the fact that Anatolian branch is the first to split off.

However, this is not a valid argument. For example, the greatest split within the Turkic language family is between Bolgharic and Common Turkic. Today the border between these branches is located in the Mid-Volga region. However, we know very well that the Turkic expansion began many thousands of kilometers further east.

Thus the fact that Anatolian branch seems to have split off first from the Indo-European unity, cannot testify for the place of this split.

The Anatolian branch can be also archaeologically satisfactorily explained from Europe.

We also know that Anatolia was full of non-IE languages; there hardly was place for Proto-Indo-European.

Moreover, Dienekes is not right concerning the vagueness of the names of the wagon parts: there are many words belonging to that semantic field, and even many words for wheel. Additionally, there are PIE words denoting to a few metals - impossible for a language of Early Neolithic.

This palaeolinguistic evidence is still stronger than any calculational dating.

eurologist said...

Average Joe:

The R1b of Europe and that of Anatolia make up two completely different, roughly equally deep stars. That is, neither one generated the other. There was no (significant) flow of R1b into Europe after LGM.

Dienekes said...

Let's play for a while that there is nothing wrong with the dating. Let's just ask: what is the basis for locating PIE in Anatolia? For all that I can see, it's just the fact that Anatolian branch is the first to split off.

The age of the split corresponds exactly to the foundation of the first Neolithic communities in Europe from Anatolia. So, if one accepts the age estimate, that's the simplest explanation, and there's no reason to invoke a separate one.


However, this is not a valid argument. For example, the greatest split within the Turkic language family is between Bolgharic and Common Turkic. Today the border between these branches is located in the Mid-Volga region. However, we know very well that the Turkic expansion began many thousands of kilometers further east.

This is not a reasonable argument, as Turkic speakers were highly mobile and expanded at a time when transportation technology was advanced and over a region that facilitated it.

The Anatolian branch can be also archaeologically satisfactorily explained from Europe.

I see absolutely no evidence for that. The best counter-argument is the virtual non-existence of haplogroup I in Anatolia and West Asia in general. A movement out of the Balkans or Central Europe would have affected the West Asian populations by bringing this typical Central European haplogroup there.

We also know that Anatolia was full of non-IE languages; there hardly was place for Proto-Indo-European.

That is incorrect. Anatolia has, throughout its history, been home to several languages and language families. It's a big place. Moreover, from the earliest times for which there is a record there are Indo-Europeans there, and there are several well-differentiated Anatolian languages there in historical times, as well as several non-Anatolian Indo-European ones. If Anatolia was not the center of the Indo-European expansion, I find it extremely surprising that Greek, Phrygian, several Anatolian languages and Iranian were all represented in it. Either Anatolia had the "Indo-European magnet", drawing diverse Indo-European peoples from a distance, or it was actually the homeland.

Additionally, there are PIE words denoting to a few metals - impossible for a language of Early Neolithic.

Copper occurs naturally, and was even used in a limited way before the Copper Age. Gold too was used long before the Copper age. Iron is also an element that would have been known to prehistoric man in its meteoritic form.

We should not confuse the emergence of full-blown metallurgy with the knowledge of metals and even their limited use. Only the latter is needed to justify having a word for them.

Average Joe said...

There was no (significant) flow of R1b into Europe after LGM.

Thanks, eurologist! I remember reading that years ago but I thought that some people were arguing that R1b had not entered Europe before the Neolithic.

Fanty said...

Hmm.

On the other site, DOES R1(a+b) roughly cover Indo European speaking countries.

I read stuff that pointed on the pattern that R1b roughly coveres the IE Languages that are called the "Centum" languages while "R1a" covered those IE languages called the "Satem" ones. (named after the word for 10, 100, 1000)

Centum/Satem map:
http://image.absoluteastronomy.com/images/encyclopediaimages/c/ce/centum_satem_map.png

R1 map:
http://wapedia.mobi/thumb/7ef4502/en/max/1440/1800/Haplogroup_R_%28Y-DNA%29.PNG?format=jpg%2Cpng%2Cgif&ctf=0?format=jpg,png,gif&loadexternal=1

Would it happen to be random?

German Dziebel said...

"The R1b of Europe and that of Anatolia make up two completely different, roughly equally deep stars. That is, neither one generated the other. There was no (significant) flow of R1b into Europe after LGM."

R1a, on the other hand, is attested in Europe from at least 4,600 years on (Eulau, Corded Ware, Haak et al. 2008), and is found at high frequencies in Central Asia, South Siberia and Western China (Uighurs), precisely where Tocharian A and B were discovered. R1a is rare in Anatolia. If R1b entered Europe after LGM from Anatolia, then R1a could be the signature of a more recent Indo-European expansion from further out east.

terryt said...

"It's actually the second worst theory of human origins, with the absolute worst being that they came from outer space..."

Created by God in the Garden of Eden has to be in the top one or two.

"The fact that the two basal branches of IE - Anatolian and Tocharian - are so widely apart from each other geographically and both are extinct with no living descendants makes it impossible to favor one geographic extreme over the other".

Why need IE have originated in one of the geaographic extremities? As Dienekes said:

"Parsimony dictates that Indo-European dispersals would not have a directional preference, and, barring natural obstacles, would radiate in all directions from the homeland".

Belenos said...

Some very amusing debate on this page. :)

I've nowt against the Anatolian hypothesis in general, but this methodology is absolutely useless, as Mr Sanchis pointed out. It rests on several patently false assumptions.

1. Rate of cognate retention can accurately date separation of languages.
2. Cognates are retained at a similar rate in different languages.
3. Cognates can always be identified, and distinguished from loans.
4. Sprachbund effects, sub/super-strata and other forms of language contact don't increase or decrease rate of cognate retention.
5. Standard dialects accurately reflect the true linguistic diversity of a language.

Belenos said...

And that's just off the top of my head, never mind the blatant inaccuracy of some of the splits, such as the division of Catalan and Occitan into differing branches when they have existed as part of the same dialect chain for most of their existance. A system that lumps Catalan with Spanish, despite the fact that the two languages clearly share many cognates due to contact effects rather than retention from common origin, will have made many, many more mistakes along the way.

eurologist said...

Dienekes, I agree y-DNA I is one big mystery.

If Anatolia was not the center of the Indo-European expansion, I find it extremely surprising that Greek, Phrygian, several Anatolian languages and Iranian were all represented in it. Either Anatolia had the "Indo-European magnet", drawing diverse Indo-European peoples from a distance, or it was actually the homeland.

Of course, one can actually take that "magnet" argument seriously and realize the evidence that IE people and languages from the outside have always set foot in Turkey. If Greek could do it (~3,000 ya?), Persian thereafter, and we had at times Romans there as well as Celts and Vikings - a lot then likely also happened before then, between ~7,000 and ~3,000 ya. Anatolia has always been at the cross roads.

Charles Nydorf said...

The chief difficulty in reconstructing the origins of the Indo-European family is that, owing to language loss the process through which the Indo-European languages emerged is hard to determine. In the absence of other information, linguists have relied on comparative reconstruction to construct a hypothetical Proto-Indo-European by working back from known Indo-European languages and assuming that they are interrelated by the shortest network of sound shifts. The assumption that the shortest possible chain is unrealistic and leads to the false conclusion that the family is only four or five thousand years old. Chronologically deeper reconstructions are possible but these involve using data from languages that are not usually considered to be related to Indo-European.

ashraf said...

Thanks for this interesting thread.

The autosomal studies too do favor a western asian homeland of proto indo-hittite (proto indo-european proper probably developped somewhere around the northwesten kazakistani region) as the indo-european speaking armenians are lacking the north european component while indo-european speaking indians lack the south european component while the west asian component is the sole one tha matches with all branches of IE.

Another support is the presence of semitic+egyptian+sumerian+kartvelian loanwords in proto indo-hittite while if we assume an european origin for pie we have to explain the total absence of loans in proto indo-european from the old european languages such as iberian, aquitanian etc...!

As for indo-european in India it could be also the result of the advancing of farmers (bringing the west asian component) and "harappan" would be an older (than indo-iranian) branch of indo-european that was later replaced by "herders/warriors" indo-iranians similary to latin replacing iliyrian (in the balkans) and celtic (in france and gaule)that did provide pre indo-european (the so called para-munda) as well as indo-european "harrapic" words into the indo-iranian languages (many cultural words with no indo-iranian etymology)
Here below an interesting linguistics paper that makes a good sum-up of the proto indo-european/hittite issue.

http://koenraadelst.bharatvani.org/print/articles/aid/urheimat.pdf


I am sorry for the typos

ashraf said...

"We also know that Anatolia was full of non-IE languages; there hardly was place for Proto-Indo-European."

On the contrary Anatolia was the home of very divergent and archaic languages of the anatolian branch (luwian, nessili, palaic, psidian, lydian, lykian, paphlagonian, pamphylian...) with 2 isolated and independant (with no sister languages in Anatolia) clearly intrucer (from the caucasus) languages i.e (very probably) west caucasic hattic and (very probably) east caucasis hurric as well as other IE branches (mitanni of indo-iranian+phrygian+armenian and perhaps other languages such as kaskis, maneans, nairi...)

The sex chromosomes origin&distribution speaks in favor of a migration from asia to europe (J,R1a,R1b,E1b,G...hg's) and not the opposite.
Finally I have the impression you omit the asian branches of IE (anatolian and indo-iranian as well as armenian) when they were attested and written down (luwian, nessili, old persian...) well before much of the IE languages of Europe, dont forget that when dealing with the IE fact , one has to remember that historically and numerically it's essentially asian rather than european and the difficulity is that there is 0 south european input and nearly 0 north european (in the case of the acrhaic ie speaking armenians and most probably in the case of the old anatolians it's 0 as well) input amongst indo-iranians.
While on the other hand the west asian component scores more than 10% in all the old indo-european biocultures of Europe (hellenes, romans) as well as eastern europeans and even amongst iberians.
Besides there was no motivation for european move into asia while the migration in the opposite way had as motivation the neolithic revolution and the introducing of farming and pastoralism to europe.

There are nowadays nearly 1,3 bln indo-iranian native speakers wich is higher than the total number of the native speakers of all the other indo-european branches grouped.

German Dziebel said...

@Terry

"Why need IE have originated in one of the geaographic extremities? As Dienekes said:

"Parsimony dictates that Indo-European dispersals would not have a directional preference, and, barring natural obstacles, would radiate in all directions from the homeland".

Well, the Anatolian homeland theory takes one of the two geographic extremities found in the basal stratum of the IE phylogeny and declares it the homeland. The western Chinese extremity represented by Tocharians is left by the wayside. I argue back that this choice is arbitrary. The Anatolian homeland prediction can't be replicated on the basis of extant languages and is therefore unverifiable. Other evidence needs to be taken into account.

Besides natural barriers, there are internal and external social/human barriers/constraints for dispersals. They may propel the founding population into a specific direction.

@Fanty

"I read stuff that pointed on the pattern that R1b roughly coveres the IE Languages that are called the "Centum" languages while "R1a" covered those IE languages called the "Satem" ones."

There does seem to be a correlation here. The problem is that the presently satem/centum division is considered by linguists to be areal, rather than genealogical. The R1a/R1b division presumably is phylogenetic. Plus if Indo-European speaking R1b folks absorbed this hg from some pre-Indo-European substrate, then the correlation may end up being coincidental. Notably, Tocharian is the only exception to the geographic association of satem languages with the East and of centum languages withe West. Tocharian is a centum language located in the easternmost extremity of the IE range. At the same time, it's heavily palatalized in ways different from the satem-style palatalization but somewhat reminiscent of other ways stop consonants are palatalized in, say, Slavic and Indic. Hittite is squarely a centum language, geographically and phonologically.

Andrew Oh-Willeke said...

We have too many calibration dates from history for this chronology to make any sense. For example, we know from historic records, to within a century, when Indo-European languages became widespread in Anatolia, and it wasn't 8,000 years ago. Archaeology can tell us when Tocharian began to be spoken in the Tarim Basin, where it probably arrived from there, and where in turn that culture derived from, and when Indo-Aryan associated cultures arrived in India.

The dates are utterly incompatible with the Anatolian hypothesis.

EliasAlucard said...

Hey Dienekes, what do you think of Gimbutas' Kurgan hypothesis and according to her, the spread of Indo-European culture over most of Europe, except the south and west of Europe, coinciding very well with the distribution of Y-DNA R1a1a?

Jean said...

The secondary products revolution is largely a failed hypothesis.

Not so. See my page on it, citing (among other things) Greenfield's 2010 review of the state of play.

Wool exists in nature now that man has bred sheep for it. It is doubtful that there was earlier a word for the undercoat of wild sheep.

Roy said...

But then one would expect an earlier branch to be closer to the source geographically, but Tocharians are further removed from Anatolia than Indo-Iranians.

Ok German, so by analogy to your geographic 'missing links' logic, the Chironomid Midges found in South Africa and New Zealand cannot be related...until we provide you proof of 'a bunch of early' common ancestors that went 'extinct' so you can join the dots...

I'm surprised Dienekes permits your childish ad hominen attacks, but then again tolerance probably correlates with intelligence.

Dienekes said...

Wool exists in nature now that man has bred sheep for it. It is doubtful that there was earlier a word for the undercoat of wild sheep.

Trouble is, that there was an early dispersal of sheep with wool of poor quality, and a later one of sheep with good quality wool

http://dienekes.blogspot.com/2009/04/retroviruses-and-origin-of-domesticated.html

Both of which started in West Asia. You can't reasonably argue that low-quality wool does not merit a word for "wool".

So, let's see, milk processing for which there is an IE word started in West Asia, wool for which there is an IE word started in West Asia...

Hey Dienekes, what do you think of Gimbutas' Kurgan hypothesis and according to her, the spread of Indo-European culture over most of Europe, except the south and west of Europe, coinciding very well with the distribution of Y-DNA R1a1a?

R1a1 could potentially be implicated in a secondary or tertirary dispersal of some Indo-European languages, namely Indo-Iranian and Balto-Slavic.

For example, we know from historic records, to within a century, when Indo-European languages became widespread in Anatolia, and it wasn't 8,000 years ago. Archaeology can tell us when Tocharian began to be spoken in the Tarim Basin, where it probably arrived from there, and where in turn that culture derived from, and when Indo-Aryan associated cultures arrived in India.

What a strange double standard. Anatolian is attested in 2,000BC, and you conclude that it was a recent arrival. Tocharian is attested in 800AD and you conclude that it was a very early arrival.

Jean said...

Trouble is, that there was an early dispersal of sheep with wool of poor quality, and a later one of sheep with good quality wool

There were two dispersals yes. But the early one was of non-woolly sheep. It's not a question of poor quality. The early sheep were kept for meat. Sheep with spin-able wool did not appear until the Copper Age. It's all on my page, with a reference to the study you mention, among others.

Dienekes said...

There were two dispersals yes. But the early one was of non-woolly sheep. It's not a question of poor quality. The early sheep were kept for meat

I see no reason to think that early Neolithic people did not use the "non-spinable wool" or that they did not have a word for it, or that PIE wool was limited only to the spin-able kind, which, anyway, originated in the Middle East.

German Dziebel said...

@Roy

"Ok German, so by analogy to your geographic 'missing links' logic, the Chironomid Midges found in South Africa and New Zealand cannot be related...until we provide you proof of 'a bunch of early' common ancestors that went 'extinct' so you can join the dots"

The comment I made was an observation on the the geography of the branches. It's just interesting that the there's stepwise geographic continuity between the "western" branches of IE (Greek, Armenian, Albanian) and Anatolian but there's no such continuity between the eastern branches (Indo-Iranian and Tocharian) and Anatolian. This observation had no critical thrust to it but it did imply that the early separation of Tocharian in the phylogeny may indicate that there were other early IE languages between Anatolia and western China.

I became openly critical of the approach only after Dienekes offhandedly attributed to me a Chinese theory of Indo-European homeland and dismissed the research of a highly respectable linguist (Johanna Nichols) that also used Tocharian in support of her Bactrian theory. Dienekes then went on to berate me with nonsensical highfalutin comments.

Nobody doubts that Hittite and Tocharian are related, so your example with Chironomid Midges misses the point. But to use the only language that you apparently can understand my qualm with the Gray Atkinson approach is that they take two extinct fossil species found in South Africa and New Zealand and assume that all the extant diversity of living species found in Europe and Asia comes from South Africa leaving the New Zealand attestation without a good explanation. So, since both the South African species and the New Zealand species are extinct species, there are only two of them but the distribution is so disjointed, one can't build an origins theory on just 2 elements in an infinite set.

"I'm surprised Dienekes permits your childish ad hominen attacks, but then again tolerance probably correlates with intelligence."

Your comment is a typical ad hominem attack. Please learn to understand other people's arguments or ask for clarifications before going on a rant. The tolerance is all mine.

German Dziebel said...

"R1a1 could potentially be implicated in a secondary or tertirary dispersal of some Indo-European languages, namely Indo-Iranian and Balto-Slavic."

R1a1 is attested in Eulau, Germany, at 2600 BC. It's also very frequent in western China, South Siberia and Central Asia - the region in which Tocharian, the second earliest branch in the IE phylogeny under discussion, was spoken. So, its range goes beyond and deeper than Indo-Iranian and Balto-Slavic, both geographically and in terms of the IE language tree.

ashraf said...

I think my post was too long to be posted here a reduplicated second part
mr Dziebel
may I have a question
do you think the west asian component is responsible of old indo-european (the satem/centum distinction is not linear and no as important as the grammatical groupings such as greco-armeno-aryan and baltoslavo-germanic grammatically motivated nodes and satemization of balto-slavic could be an areal influence from indo-iranian[the irano-slavic symbiosis of the steppes with actors like scythians/sarmatians]and of course has no relation with the R1a/R1b divide]) while the strictly connected (to the west asian component) north european component is connected to the new indo-european languages that emerged in europe as a result of the interaction between indo-european and local pre indo-european languages of europe?
As for the out of india hypothesis a major flaw is the lack of typical indian mt DNA in Europe however I dont agree that R1b should be seen (as a whole) as marker of pre (non) indı-european populations as its lack in India could be very easily explained by genetical drift such as the scenario of the proto indo-iranian trib was located in a valley east of the high ararat mountain and the mount worked as a barrier for R1b flow from further western valleys.
mr Dienekes it's a relief that you share the opinion about the "poor quality" of Anthony's book (wheel, horse...) and your discussion with mr Dziebel was highly enlightening and rich due to the remarkable intellectual quality of both
mr Alucard
There is no contradiction between the out of anatolia and the kurgan hypothesis (see works of goodenough, renfrew, bernal and gimbutas herself) I already posted about that in other forums (anthrocivitas+anthroscape)
in other words , anatolia is to be connected with the "natural" and "peaceful" colonization of old indo-european speaking farmers (see the so called northwest block as well as pleasgian, tyrsenian and minoan[also possibly cammun, ligurian, pictish and tartessian] being archaic old indo-european dialects as well as the vinca culture etc...) while the kurgan culture was (probably) connected with herders diffusing dialects such as indo-aryan and slavic.

waggg said...

"The best counter-argument is the virtual non-existence of haplogroup I in Anatolia"

I2 in Anatolia could be explained by the arrival of Phrygian I think :

http://upload.wikimedia.org/wikipedia/commons/thumb/0/06/HaplogroupI2.png/300px-HaplogroupI2.png

http://upload.wikimedia.org/wikipedia/commons/thumb/b/b6/I_Distribution.jpg/300px-I_Distribution.jpg

But what about R1a1a?
The conjunction of archeology, linguistics (early split from the IE group + Tocharian vocabulary being closer to wester IE and hittite ones than to the Indo-iranian one) and genetics in the case of the Tarim basin and south Siberian Europoid population (Chalcolithic Afanasevo culture derived from the kurgan culture of the west (J.P. Mallory)) supports better the Kurgan hypothesis IMO.
That's the closest thing to an evidence in any available theory, IMO.

Dienekes said...

This is absurd. In the Neolithic, there were plenty of human barriers to a spread, in addition to natural. Why would a population necessarily radiate in all directions? Humans populations aren't bombs, they don't explode in all directions.

On the contrary, there were not plenty of human barriers in the Neolithic. There weren't that many people around to begin with, and the Neolithic people outbred the hunter-gatherers. There is something horribly unparsimonious about a model which proposes directional migrations of Indo-Europeans from Bactria to the Atlantic, while not even a shred of population movement in the opposite direction.

Under the Anatolian hypothesis you still have to explain the emergence of one of the two most basal branches in China. There's no difference, from the parsimony perspective, which way one of the earliest founding populations went. We can choose a midpoint between the ranges of the two most basal branches, and then derive Tocharian and Anatolian from there.

We would do that only if we ignored a great deal of evidence, namely that Tocharian is attested 3 thousand years after Hittite, and there is a plethora of extant and extinct IE languages in SE Europe and West Asia, while Tocharian is an outlier.

Using the same kind of "parsimony", a future archaeologist might use Hungarian as some kind of evidence that Ugrian speakers originated in Central Europe.

ashraf said...

" IE proves to be closer to Altaic, then it would favor the Bactrian, Chinese or any other eastern homeland. If IE is closer to, sat, Afroasiatic, then it would favor the Anatolian homeland. As of now it's impossible to say. So let's not pretend we can"
I think the IE has affinities with especially the turkic and mongolic branches of IE (while "pure" altaic such as japonic and koreanic do lack those affinities) and that could be explained by sprachbund influence from the old indo-european populations of east and central asia upon the altaic and also the uralic speaking of eastern siberia and indeed there are some caucasoid input amongst the mongols wich could be a legacy of those early indo-europeans also dont forget indo-european elements in chinese,the same sprachbund seems to have taken place between an IE language(s) and the uralic branch of the uralo-siberian phylum wich is too connected racially to the north mongoloid type (nganassan/uralic component)on the other hand the afrasan-kartvelian-ie connection seems to be grammatical (inflective abluatic phyla with gender and article as well as triconsonant roots and infixes/prefixes) contrasting with the agglutinative gender+ablaut+infix+prefix+triconsonantal roots lacking structure of uralo-siberian and altaic at the end of the day the sole striking affinity between IE and altaic/uralo-siberian is the similarity between pronouns and verb endings and that could be the result of sprachbund of indo-european farmers+herders upon gathere/hunters altaic and uralo-siberians (with a consequent aquiring of pronouns-wich do have an indo-european/afrasan etymology min=who tu=female=>the "other") both ie afrasan and kartvelian share the west asian component

1/mr Dziebel do you think the internal diversity of na dene and amerind is higher than the one amongst other eurasian and african languages such as the ones within say caucaso-vasconian or even afrasan2/how to explain that there are only 2 macrophylums in the americas (na dene and amerindian)3/what about the lack of autosomal diversity as well as sex chromosomes diversity within the americas4/what about the paleoarcheological datas in the america , are they that old (I mean older than east and south africa)5/you have to explain the empty gap between the 2 major Y-DNA haplogroups (C and Q) in the americas

a said...

The geographic horseshoe band, around the Eastern Black Sea is very interesting. To the south, you have interesting remains in archeological sites, such as Gobekli Tepe(9000B.C.+/-), dug up by Klaus Schimdt. Mid range you have a elevated group of yDNA R1b1b2a(Armenians 28%+/-), some with, DYS 393-12 values putting them in with Ht-35. Elevated ydna G values around Georgia. In Europe one ancient yDNA sample G2A3 in Derenburg,Germany(5000/6000+/-). Unfortunately, perhaps all the available dna data on the Tocharians has not been given, due to any Nationalistic group or groups using the data for social unrest; perhaps some of the yDNA R1a found in the Tocharians resembles yDNA in the Northern Part of the horseshoe band around the Black Sea?

E said...

http://simon.net.nz/files/2007/12/gray_et_al_2007.pdf

In this interesting article Gray explains why this methods can be used on linguistics and addresses different problems others have pointed out regarding such works.

E said...

http://simon.net.nz/files/2007/12/gray_et_al_2007.pdf

In this material Gray explains the utility of these methods. He also addresses several concerns regarding that method also. I want to point out that he is not the only one arriving in this result

http://www3.isrl.illinois.edu/~junwang4/langev/localcopy/pdf/piazza06evolang.pdf

but also

http://books.google.com/books?id=JAbLFzVeBnMC&lpg=PA72&dq=gray%20atkinson%20albanian&pg=PA72#v=onepage&q=gray%20atkinson%20albanian&f=false

have arrived in the same results regarding PIE split.

German Dziebel said...

@ashraf

"mr Dziebel may I have a question
do you think the west asian component is responsible of old indo-european (the satem/centum distinction is not linear and no as important as the grammatical groupings such as grecoarmeno-aryan and baltoslavo-germanic grammatically motivated nodes and satemization of balto-slavic could be an areal influence from indo-iranian[the irano-slavic symbiosis of the steppes with actors like scythians/sarmatians]and of course has no relation with the R1a/R1b divide]) while the strictly connected (to the west asian component) north european component is connected to the new indo-european languages that emerged in europe as a result of the interaction between indo-european and local pre indo-european languages of europe?"

Ashraf, I'm not wedded at all to the old satem-centum division. It's very possible it's an outcome of independent innovation + diffusion. Example: Armenian is a satem language: skeSur next to Skrt svasura 'father-in-law'. But here -S- is recent and likely just a look alike of Sanskrit -S- in svaSura; if it were the same satem S, it would've been lost between vowels and we would've known it even existed.

But I do think that Tocharian forms an outgroup to both your grecoarmenoaryan and baltoslavogermanic. And Tocharian is way in the East. This is precisely where R1a reaches high frequencies. Then R1a and not R1b pop up in Europe at 2600 BC. R1b must be older, hence the West Asian component must be older, pre-dating IE and corresponding to some form of "Vasconic" substrate. It's startling that it left such an impact on Y chromosomes in Western Europe but i don't have a better explanation.

arvydas said...

Fanty said:

"I read stuff that pointed on the pattern that R1b roughly coveres the IE Languages that are called the "Centum" languages while "R1a" covered those IE languages called the "Satem" ones."

The highest R1b concentration is in Iberia and the British Isles, where Indo-European languages were NOT spoken as late as 500BC. The absolute R1b champions are Basque people and they do not speak it even now.
So for me it is hard to believe that R1b was brought in by the IE speakers, be it "centum" or "satem".

arvydas

ashraf said...

"hence the West Asian component must be older, pre-dating IE and corresponding to some form of "Vasconic" substrate. It's startling that it left such an impact on Y chromosomes in Western Europe but i don't have a better explanation"
Thanks for the answer, though the problem is that the only visible caucasoid (and hence should be connected to indo-european expansion) amongst indo-european speakers of india, iran and armenia is the west asian one (there are also some south european input amongst armenians but that south european component is totally lacking amongst indians and central asian iranians)
I think that it's the south european component that should be connected with "vasco-aquitano-ibero-pelasgian" while the west asian should be connected to a "nostratic" (kartvelian+afrasan+indo-hittite since it picks in those 3 groups[including a bit amongst afrasan speakers of eastern and northern africa]) that diverged later to afrasan, kartvelian& old indo-hittite.
For the kartvelian-afrasan-indohittite affinities I read about them in martin bernal's book quoting various linguists as well as bernard sergeant.
To have an idea about some of those affinities please see the thread below
http://anthrocivitas.net/forum/showthread.php?t=7697&page=2

As for the north european component it seems that it originated in the caucasus and was responsible of the colonization of eastern europe after the retreat of ice sheets.
The south european one seems connected with the ice age refugium in europe and there is a possible connection of ibero-aquitanian with caucasic languages (however a moribund link because of the very deep age of the breakup of the supposed vasco-caucasic macrophylum)here a below a sum-up of those speculations in a graph
http://anthrocivitas.net/forum/image.php?s=cbc61065285cc8cad795bd98b2dc809b&u=781&type=sigpic&dateline=1296937096

Average Joe said...

The absolute R1b champions are Basque people and they do not speak it even now.

But couldn't R1b in the Basque population be the result of gene flow from surrounding Iberian populations? We have no idea how long R1b has been in the Basque population.

German Dziebel said...

@ashraf

"the west asian should be connected to a "nostratic" (kartvelian+afrasan+indo-hittite..."

Just a note: Afroasiatic (if it's the same as your "afrasan") is now virtually excluded from the Nostratic unity by most Russian Nostraticists (loanwords into IE is a different matter, of course). Afroasiatic is still part of the Euroasiatic superfamily but this one has even fewer supporters. There's no other "in" for IE into West Asia in the Nostratic superfamily.

"I think the IE has affinities with especially the turkic and mongolic branches of IE (while "pure" altaic such as japonic and koreanic do lack those affinities.."

Nostraticists see IE as closely related to Uralic and Altaic before Kartvelian, Dravidian, etc. Korean and Japanese are macro-Altaic languages, so you got things a bit reversed here. But I guess we're looking at the crux of the matter. If IE is next-of-kin to Afroasiatic, then the Anatolian homeland becomes likely, if the similarities between IE and Afroasiatic come from areal influence and convergence, then the Uralo-Altaic pole will make an eastern origin for IE likely. I guess this is where your "ablautive vs. pronoun/verb ending" dichotomy comes in. Judging from the recently established Ket-Na-Dene link, systematic matches in the pronominal and verbal paradigms are a secure sign of linguistic kinship.

As for Kartvelian, I know that the Basque-Kartvelian linguistic link has been dismissed, but in my core field of expertise, namely kinship systems and terminologies Kartvelian (especially Svan) and Basque share unique similarities (see Dziebel, "The Genius of Kinship," 325-326). http://books.google.com/books?id=1yo14jYY3pYC&pg=PA326&lpg=PA326&dq=genius+kinship+kartvelian&source=bl&ots=1QOGGNBCxN&sig=r81HaGj06O8Ji48zr6LjKoA003s&hl=en&ei=z7WjTc7oIKnd0QGwmeWECQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CBUQ6AEwAA#v=onepage&q&f=false

"1/mr Dziebel do you think the internal diversity of na dene and amerind is higher than the one amongst other eurasian and african languages such as the ones within say caucaso-vasconian or even afrasan2/how to explain that there are only 2 macrophylums in the americas (na dene and amerindian)3/what about the lack of autosomal diversity as well as sex chromosomes diversity within the americas4/what about the paleoarcheological datas in the america , are they that old (I mean older than east and south africa)5/you have to explain the empty gap between the 2 major Y-DNA haplogroups (C and Q) in the americas"

This is a huge question and I'm not sure the comments section is the best place to discuss it. Two quick things; the diversity within "Amerind" is exceptional and unmatched anywhere in the world with a possible exception of Papua New Guinea. Y-DNA: I don't think hg Q is as downstream as phylogenies portray it. I think it was misallocated to the bottom of trees but in fact P (Q, R) must be of the same depth as C-F and D-E. (Otherwise, we would've seen hgs N and O in the Americas.) Intragroup genetic diversity is a function of effective population size. Amerindian populations have always been demographically depressed and isolated (their intergroup diversity values are the highest in the world). This is also key to the paucity of archaeological sites in the Americas, although thanks to Monte verde and Buttermilk we know for a fact that Amerindians were present south of the ice sheets from where they re-expanded north with the retreat of the glaciers (Na-Dene?). Amerindians retain Mid-Pleistocene demographic parameters better than the expanded Old World populations. In Africa, Hadza is a trace of this ancient demography of the Amerindian kind. Other populations are just expanded and intermixed, hence intragroup diversity is high.

German Dziebel said...

@Jean

"Wool exists in nature now that man has bred sheep for it. It is doubtful that there was earlier a word for the undercoat of wild sheep."

It's interesting that the IE term for 'wool' (*ulH2na) is attested in ALL branches of IE. Not only that the root of the word is the same throughout the IE area but the morphology is the same, too. There's slight phonomorphological variation suggesting that the word is not just a wandering word. http://books.google.com/books?id=tzU3RIV2BWIC&pg=PA648&lpg=PA648&dq=indo-european+wool&source=bl&ots=wVj_06a87G&sig=jaf3DbCbgyq74NjhzlyNfxQfXd0&hl=en&ei=FlukTYD4Oq6y0QGy7eGACQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CB0Q6AEwAA#v=onepage&q=indo-european%20wool&f=false

Hittite is no exception here. The word for 'wool' in Hittite is *Hulana (< *ulH2na). The only branch that doesn't have the reflex of this word is Tocharian. Toch B word for 'wool' is yok, and apparently it's not a borrowing but a rare IE word. It's possible therefore that Tocharian split from the rest of IE languages prior to the invention of wool and the word for it.

ashraf said...

The russian school was accused of being poltically biased i.e they included altaic and uralic in nostratic to legitimize the russian colonization of the altaic and uralic speaking lands.
I think the grammar is way more important when linking languages than pronouns, and the pronouns and verb endings of uralic and altaic could be the result of sprachbund by expanding west asians (such as scythians) and one has to remember that the first attested altaic or uralic languages was the 8 th century orhon turkic and before that date many water should have been passed under the bridge, afrasan, kartvelian and dravidian are included in nostratic but not in eurasiatic, if you read "aron dolgopolsky's nostratic dictionnary" you could notice that there are matches between afrasan-kartvelian-IE pronouns as well and that the core vocabulary shared by the three is higher than the ones with altaic-uralic.
The second difficulity is that altaic and uralic are connected with the mongoloid type i.e there is not a sahred genetical component between the 3 while we have the west asian caucasoid component being shared by afrasan-kartvelian-IE.
As for the internal diversity of amerind, are you sure amerind is more diverse that phyla such as nilo-saharan, caucasic?
2/still there are only 2 languages families in the americas with one being present also in asia
3/there are even nowadays many language isolates in africa such as kxa, tuu and hadza and very likely before the language shift of many african groups to the dominant afrasan/nilo-saharan/niger-kongo there must be many other highly diversified language isolates.
There is not even a taxonomic dichtomy of the kind of altaic/macro-altaic or uralic/uralo-siberian.
How, then could amerindian be so diverse and afaik most of the amerindian languages have the same agglutinative morphology yet in africa you have languages of the same family but of different structure (see inflective semitic and agglutinative cushitic)
Finally one has to remember that central asia was populated by western asian advancing folks[most likely farmers and pastoralists] (as shown by the autosomal studies)so central asia cannot be the source but rather the receiver.
It's obvious that the overcrowded indus valley bioculture could not be populated and "linguistically dominated" by few central asian nomads(and anyway there is no central asian input amongst indians but chiefly south asian+west asian)it looks obvious that IE in India was the result of natural colonization by farmers advancing from the agriculture cradle in west asia and there was an intermingling with the locals (south asian component) as it's shown by autosomal studies.

waggg said...

@ German Dziebel : R1a1 is attested in Eulau, Germany, at 2600 BC. It's also very frequent in western China, South Siberia and Central Asia - the region in which Tocharian, the second earliest branch in the IE phylogeny under discussion, was spoken. So, its range goes beyond and deeper than Indo-Iranian and Balto-Slavic, both geographically and in terms of the IE language tree.

Indeed.
And IMO there are indirect prooves that the europoid chalcolithic R1a1a population of south siberia (coming from the west) was an (non satem) IE people.

There seem to be an IE cattle trail in the south Siberian/Mongolian area :

The mongolian cattle has some european input (and interestingly Turkic öküz (ox) seems derived from a typical IE root : Tocharian B okso, Danish Okse, etc... and Turkic is thought to come from roughly the south sibera/Mongolia area (the first attested Turkic inscriptions are found in the north of Mongolia).

http://dienekes.blogspot.com/2010/07/genetic-structure-of-cattle-in-eurasia.html


And the Japanese cattle also has a quite important european input (in this case too, we can find an very likely IE linguistic connection as Ainu and Japanese dialects got peko and beko (*) which are pretty similar to PIE *peku (e.g. latin "pecus", etc...))).

“Mitochondrial DNA variation and evolution of Japanese black cattle (Bos taurus)”"

(*) http://en.wikipedia.org/wiki/List_of_Ainu_terms#P

It makes sense if the last migration in Japan came from south Siberia via Korea.


The best explanation is that the europoid chalcolithic R1a1a pastoralist population of south Siberia was speaking an IE language (non satem). who else in south Siberia/Mongolia? What other archeological trail? what other haplogroup than R1a1a as vector from the west in this place?

Besides it's very likely that the R1a1a Xiaohe mummies in Xinjiang came from south Siberia (Chunxiang Li et al, 2010) so it's also quite likely (or at least quite credible) that "proto-Tocharian" was spoken in chalcolithic Afanasevo culture of south Siberia.
According to J. P. Mallory, they were clearly derived from the population from the south Russia (many similarities with the Kurgan culture of the west attested in archeology).
Hence their language came from the north-east of the black sea, not Bactria (anyway neither their europoid morphology neither their mtDNA haplogroups leave doubts (ancient south Siberian and Tarim R1a1a haplotypes and mtDNA lineages also have matches in modern Europe)).

German Dziebel said...

@ashraf

"The russian school was accused of being poltically biased i.e they included altaic and uralic in nostratic to legitimize the russian colonization of the altaic and uralic speaking lands."

Nonsense.

"I think the grammar is way more important when linking languages than pronouns, and the pronouns and verb endings of uralic and altaic could be the result of sprachbund"

Grammar by itself is not how linguistic kinship is established. Grammar is driven by independent invention. It's the lexical and phonetic material that codes for those grammatical patterns (and pronouns and verb endings are part of grammar) that matters. Agglutination, etc. has no phylogenetic value. It comes in goes as it pleases.

"if you read "aron dolgopolsky's nostratic dictionnary" you could notice that there are matches between afrasan-kartvelian-IE pronouns"

yes, Dolgopolsky is the only heritage member of the Russian Nostratic School that believes that Afroasiatic is related to other Nostratic languages. For everybody else (Dybo, Starostin, etc.) they are connected areally.

"altaic and uralic are connected with the mongoloid type i.e there is not a sahred genetical component between the 3 while we have the west asian caucasoid component being shared by afrasan-kartvelian-IE."

Uralic is an independent "race" (see the tertium quid theory of Uralic). It's not Mongoloid, Caucasoid or a mix of both. Afroasiatic has languages that are spoken by Black Africans and languages that are spoken by Caucasoids. So, a language family can easily cut across racial divides, either through language shift or because it predates the phenotypical split.

"How, then could amerindian be so diverse and afaik most of the amerindian languages have the same agglutinative morphology yet in africa you have languages of the same family but of different structure (see inflective semitic and agglutinative cushitic)"

Amerindian languages show phenomenal grammatical differentiation: agglutinative, flexive, incorporating, isolating, polysynthetic, active, ergative, accusative - what have you. It's the only continent that attests to all logically possible version of word order (SOV, VSO, etc.)

"As for the internal diversity of amerind, are you sure amerind is more diverse that phyla such as nilo-saharan, caucasic?"

I'm absolutely positive that "Amerind" is more diverse than the other phyla. It's so diverse that it's not even considered a phyla or a family. The whole "Amerind" grouping was debunked and the most recent volume from The Handbook of American Indian Languages - Linguistics - doesn't even mention Amerind. There are 140 (!) unique language families in the Americas. This is 2/3 of global linguistic variation. In Africa, we have 20 families at best.

"2/still there are only 2 languages families in the americas with one being present also in asia"

See above regarding the true number of language families in the Americas. The two most northerly families - Na-Dene and Eskimo-Aleut - have recognized relatives in Asia. There are 138 more of them further down south in the New World.

Andrew Oh-Willeke said...

"On the contrary Anatolia was the home of very divergent and archaic languages of the anatolian branch (luwian, nessili, palaic, psidian, lydian, lykian, paphlagonian, pamphylian...) with 2 isolated and independant (with no sister languages in Anatolia) clearly intrucer (from the caucasus) languages i.e (very probably) west caucasic hattic and (very probably) east caucasis hurric as well as other IE branches (mitanni of indo-iranian+phrygian+armenian and perhaps other languages such as kaskis, maneans, nairi...)"

There were lots of IE languages in Antaolia post-Hittite when the empire crumbled, a lot like there were lots of Romance languages that emerged when the Roman empire crumbled.

The case that IE languages were widespread pre-Hittite is far harder to make. When Hittite is first attested in Anatolia ca. 2000 BCE, it is present in just one or two cities and rapidly expanding with the empire into an overwhelmingly Hattic and Hurrian speaking area. There might have been a few small communities that spoke kindred IE languages in Anatolia when Hittite was confined to just a couple of cities, but these were clearly the new kids on the block at that point.

This expansion is close in time to the replacement of pre-IE people in Greece with Greeks, and with the arrival of cultural coordinates of IE languages like cremation appearing in Greece, the Balkans, Anatolia and North India. Where ever proto-IE was spoken, it was virtually absent from Western Europe, Greece, Anatolia, Persian and greater India until the early Bronze Age. Mitanni appears to the East of the Hittite Empire in a specifically Indo-Aryan dialect just when one would expect it if it was expanding from North India over the past several hundred years, and is consistent with modern Persian and Hindi both having common descent from an earlier Indo-Iranian language.

The historical and archaeological evidence is one of IE suddenly bursting onto the scene and becoming rapidly dominant over a very broad geographic area all at once about 4000 years ago, probably in connection with the climate disaster that struck in SW Asia and Northern India at about that time.

Tracing the cultural antecedents of the Tocharians (and assuming that those cultural antecedents spoke an IE language), suggests an origin for PIE in the European Steppe well before the big breakout for IE languages (ca. 3500BCE to 4000 BCE), but as a quite minor language with little lasting impact until a much later in time breakout.

There isn't cultural continuity between the Danubian Neolithic and the proto-IE Kurgans. It is much more reasonable to expect than that Danubian Neolithics spoke a language more like Hattic or Hurrian or Lemnian or Minoan or Sumerian, none of which are particularly similar to IE.

Andrew Oh-Willeke said...

"What a strange double standard. Anatolian is attested in 2,000BC, and you conclude that it was a recent arrival. Tocharian is attested in 800AD and you conclude that it was a very early arrival."

The early attestation of Anatolian then suggest that Hittite was confined to a couple of cities and that that they were a decided minority in Anatolia at the time. Almost all of the Hittite expansion from its second city onward was documented in historical writing. They were probably a recently arrived tribe exiled from elsewhere to Anatolia for reasons unknown that decided to set down roots with a vengence.

The Tocharian language speaking people had clear cultural continuity in archaeology back to about 2000 BCE and have clear cultural antecedents to the near North of the Tarim basin that go back another 1000 plus years, and those peoples in turn have apparent antecedents in the European Steppe.

Genetically, the Tocharians at least had Y-DNA hg R1a suggesting that they were probably patriline descendants of Neolithic farmers. But, ethnogenesis of PIE probably happened much later than the early Neolithic.

terryt said...

"in fact P (Q, R) must be of the same depth as C-F and D-E. (Otherwise, we would've seen hgs N and O in the Americas.)"

Not necessarily so. Perhaps N and O didn't begin to move north of latitude 45 North until after Q and C had gone past through that region. That is certainly what I have assumed for some time now.

ashraf said...

"IndoEuropeans originated in the Pontic-Caspian Steppe"Those steppe indo-europeans themselves originated in west asia according to maria gimbutas herself see 2 papers below
http://koenraadelst.bharatvani.org/print/articles/aid/urheimat.pdf http://anthrocivitas.net/forum/showthread.php?p=86523
We all berrnal+renfrew+gimbutas+ivanov
accept the origin of Indo-Hittite in Anatolia.I agree with Renfrew that the initial move was made much earlier with the spread of agriculture.I see the migrants’ language not as Proto-Greek but as a branch of Indo-Hittite.Peoples speaking forms of this language spread north to create the Neolithic civilizations of the Sixth and Fifth Millennia BCE in the Balkans...people from these agricultural civilizations on the edge of the steppe developed techniques of nomadism. From this mixed agricultural and nomadic population that spoke Indo-Hittite the Kurgan culture formed and Indo-European, in the narrow sense...I accept that the Kurgan culture and Indo-European languages spread out from the steppe.What Gamkrelidze and Ivanov call the Ancient European Dialects(Celtic,Italic,Illyrian,Germanic,Baltoslavic and Tocharian)derived from northern dialects and migrated earlier, while the Indo-Aryan(Armenian and Greek) came from the southern.
As for kartvelian-afrasan-IE genetical (and not sprachbundic as it's the case with uralic and altaic) connection please see the 2 papers below
http://anthrocivitas.net/forum/showthread.php?p=75413
http://koenraadelst.bharatvani.org/print/articles/aid/urheimat.pdf
Semitic (and by implication also the Chadic, Kushitic and Hamitic branches of the Afro-Asiatic family, assumed to be the result of a pre-4th-millennium immigration of early agriculturists from West Asia into North Africa) is suspected to spring from a common ancestor with IE, even by scholars skeptical of Nostratic adventures. The commonality of some elementary lexical items is striking, and includes the numerals 6 and 7 (Hebrew shisha, shiva, Arabic sitta, sab'a, conceivably borrowed at the time when counting was extended beyond the fingers of a single hand for the first time), arguably even all the first seven numerals.Even more remarkable are the common fundamental grammatical traits, which indicate a common genetic origin rather than an influence from the one language family on the other. Semitic,like IE, has grammatically functional vowel changes, grammatical gender,declension, conjugational categories including participles and medial and passive modes, and a range of phonemes which in Proto-Semitic was almost entirely in common with PIE, even more so if we assume PIE laryngeals to match Semitic aleph, he and 'ayn. Many of these grammatical elements are shared only by Semitic (or Afro-Asiatic) and IE, setting them off as a pair against all other language families. If any language family has a chance of being the sister of the IE family, it is Semitic. One way to imagine how Semitic and IE went their separate ways has been offered by Bernard Sergent, who is strongly convinced of the two families' common origin. He combines the linguistic evidence with archaeological and anthropological indications that the (supposedly PIE-speaking) Kurgan people in the North-Caspian area of ca. 4000 BC came from the southeast, a finding which might otherwise be cited in support of their Indian origin. Thus, the Kurgan people's typical grain was millet, not the rye and wheat cultivated by the Old Europeans, and in ca. 5000 BC, millet had been cultivated in what is now Turkmenistan (it apparently originates in China),particularly in the mesolithic culture of Jebel. From there on, the archaeological traces become really tenuous, but Sergent claims to discern a link with the Zarzian culture of Kurdistan 10,000 to 8500 BC. Short, he suggests that the Kurgan people had come along
the eastern coast of the Caspian Sea, not from the southeast (India) but the southwest, in or near Mesopotamia, where PIE may have had a common homeland with Semitic

ashraf said...

part 2
As for the similarity of the pronouns in afrasan/kartvelian/IE please see pp 28-34 of Dolgopolsky's book here below
http://www.dspace.cam.ac.uk/bitstream/1810/196512/46/01ND_Intro_7-84.pdf
As for the IE influence on Uralic and the connection of the later with the mongoloid type please see the paper below
www.sgr.fi/sust/sust258/sust258_janhunen.pdf
in other words tongue affinity+genome affinity=>genetical link but tongue affinity without genome affinity=>sprachbund(both angolans and lusitanians are portuguesic)
IE affinity with uralic and altaic in the pronouns and verb endings should be explained by an IE sprachbund perhaps from a scythian origin that resulted in uralic+turkic+mongolic adopting IE pronouns and verb endings(before that they could have been lacking pronouns and verb endings, it's important to notice that pure altaic languages as japanese lacks verb endings and has different pronouns than IE)
proto turks proto mongols and proto uralics borrowed iranian script and religious system and iranian cultural elements(iranian ahuramazda borrowed by mongolians-indo-iranian deity druga borrowed by turkic tengri as well as soghdian wolf myth borrowed by turkic which did even retain the iranian names of the myths such as ashina, kurt, bori and bortechine)
So it should not be difficult to imagine that they borrowed iranian pronouns and verb endings as well.
As for the indo-iranian, it seems they came from west asia and not europe or central asia as shown by autosomal studies where indo-iranians appear to be a mixture of west asians+south asians(for indo-aryans)and west asian+southwest asian(for iranians)
One had to remember that indo-iranians were first attested in west asia(mitannis and scythians/ashkuz) also kassites and manneans could have been indo-iranian as well but sadly since no document written in kassic or maneanic was preserved to our days we cannot know much about that.
Finally, it's interesting that non IE speaking basques and originally non IE speaking east iberians(who were speaking iberian a language-very likely-connected to basque/aquitanian) as well as uralic speaking finns do lack the west asian component wich strengthens the possibility that west asian component in Europe was connected with the introduction of the indo-european languages into Europe.
As for Dravidian, its affinity with IE is to be explained with sprachbund too; nowadays french is very close to french creole spoken by caribeans but we know very well that there is no original common ancestry between french peoples and caribeans and a similar case should have occured between dravidic speaking south asians and ie speaking west asians.
As for kartvelians having affinities with both caucasic and nostratic=>it could be expalained by a caucasic superstratum (hattic?) on kartvelian.
"Uralic is an independent "race" (see the tertium quid theory of Uralic). It's not Mongoloid, Caucasoid or a mix of both. Afroasiatic has languages that are spoken by Black Africans and languages that are spoken by Caucasoids. So, a language family can easily cut across racial divides, either through language shift or because it predates the phenotypical split."
No language family could be as old as the caucasoid-mongoloid-kongoid shift,Uralic is north mongoloid from the uralic/nganassan type,afrasan in africa was introduced by southwestasians in the 4th millenium BC as its shown by the caucasoid input(southwest asian component)amongst eastern africans(20 to 50%)and north africans(25%to 50%)=>it's language shift please see below for the afrasan homeland
http://anthrocivitas.net/forum/showthread.php?p=114735
and here below for the IE homeland
http://anthrocivitas.net/forum/showthread.php?p=105442

Dienekes said...

The early attestation of Anatolian then suggest that Hittite was confined to a couple of cities and that that they were a decided minority in Anatolia at the time.

Hittite is not the only Anatolian language that is early attested in Anatolia. Luwian and Palaic are also attested, and a millennium later the entirety of West Asia Minor spoke Anatolian languages (minus the Greeks and Phrygians).

Proponents of the Kurgan theory have a hard time explaining how Anatolian arrived in Anatolia. It takes a lot of weird migration routes and a lot of hide-and-seek to account for the fact that it is so distantly related to languages from which it supposedly branched out late in prehistory.

ashraf said...

Not only Hittite(nessili) and Luwili but dozens of Anatolian Indo-European languages all over Anatolia (and very diversified) and according to own Hittite accounts they came from southeast near a big lake(van lake??), besides according to gamkrelidze and ivanonv the toponyms of the area around eastern anatolia are the only proto indo-european ones (i.e near to the constructed proto indo-european forms)
here below a non complete list of the indo-european languages of the anatolian branch:
nessili, luwian (2 dialects), palaic, lydian, lykian, pamphylian, likaonian, isaurian, psidian, carian...
Here below a map of some of those languages
http://dnghu.org/anatolian-languages-map.jpg
As you can notice the Anatolic speaking folks were confined to south and eastern anatolia with no opening on the black sea nor the aegean sea wich is in line with the hittite account that they came to anatolia from the south-east from a land with a big lake (urmia, van??)

German Dziebel said...

@ashraf

"No language family could be as old as the caucasoid-mongoloid-kongoid shift"

You quote Dolgopolsky all the time, and he believes in Nostratic as a genealogical grouping, not a Sprachbund. Dolgopolsky's Nostratic is a "family" that has Caucasoid, Black African, Uralic and Mongoloid components to it. Needless to say, of the theory proves right, then this language family would be old enough to transcend some of the racial divides.

"in other words tongue affinity+genome affinity=>genetical link but tongue affinity without genome affinity=>sprachbund(both angolans and lusitanians are portuguesic)"

There's a reverse process, too, namely language preserve old kinship but gene flow obscures the original genetic connection between pairs of language branches. Mon-Khmer and Munda are part of the same language family, but their mtDNAs are distinct because Munda absorbed lots of genes in India. Only Y-DNA preserves the original connection between MK and Munda as well as their languages. Ket and Na-Dene are related linguistically but genetically there's no similar link. It's likely that the Ket genetic makeup was changed through gene flow with neighbors. Burusho speak and isolated language but genetically they are just like their neighbors. This again means that they are saturated with later genes but speak an ancient language.

We just tend to see prehistory in the light of population movements and language shifts that occurred as result of the most recent worldwide European colonization. But this is an exceptional event enabled by cultural and demographic factors that weren't in place in older times.

"So it should not be difficult to imagine that they borrowed iranian pronouns and verb endings as well."

I'm sympathetic with your approach to tease apart areal and genetic traces but I don't think we can attribute Uralic and Altaic pronouns and verb endings to Indo-Iranians. Same for Dravidians. Later borrowings are usually easy to detect, but all Nostraticists and Eurasiaticists would tell you that those grammatically features are genetically linked, not areally. That's the whole point of the Nostratic/Eurasiatic theory. Gene flow and linguistic borrowing can be attributed to the IE-Afroasiatic interface instead. That's why you get a West Asian component in some IE-speaking populations but not in others.

Thanks for all the references, Ashraf.

German Dziebel said...

@Terry

"Perhaps N and O didn't begin to move north of latitude 45 North until after Q and C had gone past through that region."

Q and C are currently on the two opposite ends of the Y-DNA phylogeny. It doesn't make sense that they both ended up north of latitude 45 North, a region that has never been a "magnet" for populations. At the same time N and O that are presumably closer together with hg Q phylogenetically, hadn't been around it geographically when the New World was colonized but once it was settled they poured in and displaced it. Plus Q is a sister clade of R and, together, they are spread across three continents (America, Eurasia and Africa) suggesting great antiquity. Ancient hgs such as C and D-E typically show a phenomenal geographic spread. We see the same thing with "Neanderthal-derived" X chromosome hg B006. It's found in America, Australia and Europe. If we upgrade Y-DNA hg P to the level of CDE then everything will fall into places nicely. And I don't mean "out of America" but just the concordance between phylogeny and geography. Molecular clock just ticks faster for some clades and slower for others. And some nodes such as K and F are just artificial place-holders.

@Dienekes

"Proponents of the Kurgan theory have a hard time explaining how Anatolian arrived in Anatolia. It takes a lot of weird migration routes and a lot of hide-and-seek to account for the fact that it is so distantly related to languages from which it supposedly branched out late in prehistory."

"Distantly related" is a tricky concept. First of all, Anatolian is an extinct branch, so we don't know what it descendant languages would've looked like if they survived. The only reason Armenian is part of the IE family is because of the Old Armenian attestation. Without it, modern Armenian and modern IE languages would be hardly recognizable as part of the same family. Consequently, other modern languages that aren't IE may not be IE because of the lack of a 3,500 year old attestation.

On the other hand, as in the case of the IE term for 'wool' above (or let's take the word watar 'water'), Hittite matches other IE languages literally. If we take such part of the basic IE vocabulary as kinship terms, Anatolian deviates markedly from other IE languages in some ways (anna 'mother' next to IE *meHter) but in others it preserves the very exact morphology of typical IE terms (Lyk. cbatra [cb < dw], Luw duuttari 'daughter'). This suggests that Anatolian may have shifted away from the common IE heritage, rather than preserved the earlier stage of IE.

azarov_dmitry said...

Great research. It completely proves the kurgan theory and the fact that the Hittite is NOT EI language. This research is consistent with the theory of the Indo-Hittite language according to which the Anatolian evolved from the Indo-Hittite (Indo-Hittite->Pre-Anatolian->Anatolian). While IE evolved from the Indo-Hittite through PIE (the Indo-Hittite PIE->IE). That perfectly explains huge differences between the Hittite and other IE languages. So that we can conclude that in fact this research sets the timing of split between Pre-Anatolian and PIE and so the oldest attested IE language is the Tocharian.

Jaska said...

Dienekes:
“The age of the split corresponds exactly to the foundation of the first Neolithic communities in Europe from Anatolia. So, if one accepts the age estimate, that's the simplest explanation, and there's no reason to invoke a separate one.”

Yes there is: Anatolian homeland cannot explain the PIE vocabulary as well as Ukrainian homeland.

Dienekes:
“This is not a reasonable argument, as Turkic speakers were highly mobile and expanded at a time when transportation technology was advanced and over a region that facilitated it. “

Still there is no reason to assume that PIE was spoken in the Anatolia just because the Anatolian branch was the first to split off. It would also be more economical to locat the homeland closer to the centre of gravity of the language family.

Jaska:
“The Anatolian branch can be also archaeologically satisfactorily explained from Europe.”
Dienekes:
“I see absolutely no evidence for that. The best counter-argument is the virtual non-existence of haplogroup I in Anatolia and West Asia in general. A movement out of the Balkans or Central Europe would have affected the West Asian populations by bringing this typical Central European haplogroup there.“

If you would read the books of Mallory etc. you could see the archaeological explanations: there are very well fitting expansions to Anatolia. Absence of haplogroup I cannot be a counter-argument, as there is no reason to believe that there were any I in Ukraine at that time.

Jaska:
“We also know that Anatolia was full of non-IE languages; there hardly was place for Proto-Indo-European.”
Dienekes:
“That is incorrect. Anatolia has, throughout its history, been home to several languages and language families. It's a big place. Moreover, from the earliest times for which there is a record there are Indo-Europeans there, and there are several well-differentiated Anatolian languages there in historical times, as well as several non-Anatolian Indo-European ones. If Anatolia was not the center of the Indo-European expansion, I find it extremely surprising that Greek, Phrygian, several Anatolian languages and Iranian were all represented in it. Either Anatolia had the "Indo-European magnet", drawing diverse Indo-European peoples from a distance, or it was actually the homeland.”

We know very well that Greek came to Anatolia from the west and Indo-Aryan came to Anatolia from the east. Nobody believes that Greek or Aryan were born in Anatolia!

The only languages, from which we have no firm evidence outside Anatolia, are the Anatolian languages. But even in these languages we see traits of massive substrate influence, which points to the conclusion that even the Anatolian languages are not original in Anatolia.

If Anatolia was earlier Indo-European, there should not be so strong foreign influence in the Anatolian languages.

Andrew Oh-Willeke said...

"Hittite is not the only Anatolian language that is early attested in Anatolia. Luwian and Palaic are also attested, and a millennium later the entirety of West Asia Minor spoke Anatolian languages (minus the Greeks and Phrygians).

Proponents of the Kurgan theory have a hard time explaining how Anatolian arrived in Anatolia. It takes a lot of weird migration routes and a lot of hide-and-seek to account for the fact that it is so distantly related to languages from which it supposedly branched out late in prehistory."

Jaska has good points which I mostly won't repeat. But, the most important point is that the apparently distant relationships can be explained by varying degrees of pressure from competing languages in the region, both substrate influences and areal influences.

The Norse v. Icelandic example is on point. A linguistically isolated community like Iceland is going to evolve very slowly, while a linguistic community like the Norse with all sorts of neighbors is going to evolve much more rapidly because of the substrate and areal influences that put pressure on it, and most of that divergence is going to happen sooner rather than later in punctuated linguistic evolution.

The Myceneans whose Indo-European language becomes ancient Greek around 1600 BCE upon their arrival and conquest of the Pelasgians and Minoans are influenced by a non-Indo-European substrate strong enough to assert a continuation of the Minoan Linear A script into Linear B, to be incorporated into Indo-European polytheism rather than being replaced by it, and to exist as a subculture within ancient Greece for hundreds of years after the initial conquest.

There is no good reason to assume that Luwian and Palaic had any longer history in Western Anatolia than Hittite. They could have been separate tribes that were part of the same Hittite migration, or could have been part of the same migration as the Myceneans (if indeed they were even separate), but evolved into Anatolian languages due to Hattic substrate influences and areal influences.

Andrew Oh-Willeke said...

Continued:


"a millennium later the entirety of West Asia Minor spoke Anatolian languages (minus the Greeks and Phrygians)."

A eight hundred years of rule of the entire region by the Hittite empire will do that. Successive waves of Celts and Romans had a similar impact in a similar time frame in all of non-Basque Iberia. This is about the same time frame in which Italic languages went from being non-existing minor colonies in Italy to wiping out Etruscan as the last vestigate of the Villanovian substrate there. Akkadian completely supplanted Sumerian outside religous texts in that time frame. Turkish became the language of Anatolia faster than that.

The official language of the Hittite empire, that over time became the popular language of the empire prior to its collapse about 800 years later, was not an amalgam of different Anatolian languages. It is was the language of a single Anatolian city ca. 2200 BCE that bit by bit conquered a larger and larger empire until its sphere of influence extended to include almost all of Anatolia and reached down to the Northern Levant until it bumped up against the Egyptians. The story of Anatolian languages is a lot like the story of English which started as the language of a tiny Friscan dialect and came to be spoken all over the world, or of Arabic, or of Bantu, or of Roman, or of Han Chinese, each of which have similar histories.

For the first few hundred years, first as a minority people in a foreign land and then as a thin superstrate ruling elite (transcribed in a writing system adapted from one previously used for Semitic languages), this Hittite dialect was subject to much greater outside influence than, for example, Tocharian. Moreover, the ruling class that set the standard was actively intermarrying and conducting diplomacy with Hattic, Hurrian, Mittani, Akkadian and Egyptian speaking peoples, to a greater extent than the Indo-Iranians or the Mycenians. They also borrowed a lot of metallurgical technology from the Caucasians. Little wonder then, that Hittite, in a linguistic cross roads had less lexical similarity to other IE languages and more borrowing from languages now lost.

One also can generally expect that the rate of lexical change is going to be much slower once these languages are literary than it is when these are purely oral languages. The transition took place after these languages had already differentiated. The other examples given "Semitic, Austronesian, Melanesian, the languages of the Sahul, and Arawak" mostly didn't have this transition to seriously interfere with dating accuracy, and mostly didn't have strong competing language families to "fight back" and influence the measured language strongly and in a way that would impact different branches at a different rate from each other.

Dienekes said...

The Myceneans whose Indo-European language becomes ancient Greek around 1600 BCE upon their arrival and conquest of the Pelasgians and Minoans

That sentence has so many fallacies in it.

1. There is no evidence that the Mycenaeans "arrived" in 1600BC
2. There is no evidence that Greek "emerged" in 1600BC from Indo-European. On the contrary, Mycenaean Greek is already fully recognizable Greek, so it must've been Greek for quite some time before, not to mention that it was also dialectal Greek, so its divergence from Proto-Greek must be pushed into the past, certainly in the 3rd millennium BC at the latest.
3. There was no conquest of "Pelasgians", that's simply outdated philology.

The official language of the Hittite empire, that over time became the popular language of the empire prior to its collapse about 800 years later, was not an amalgam of different Anatolian languages.


If Hittite was the only Anatolian language in Anatolia, you'd have a point.

terryt said...

"Q and C are currently on the two opposite ends of the Y-DNA phylogeny. It doesn't make sense that they both ended up north of latitude 45 North, a region that has never been a 'magnet' for populations".


But some sort of humans have been there for a very long time. And I don't accept for a moment that they entered America together.

"At the same time N and O that are presumably closer together with hg Q phylogenetically, hadn't been around it geographically when the New World was colonized but once it was settled they poured in and displaced it".

I'm quite sure that is so. That explains the gap between Q in the Ket/Selkup people and the Americans. But NO and Q had taken completely different routes into North/Central Asia.

"Plus Q is a sister clade of R and, together, they are spread across three continents (America, Eurasia and Africa) suggesting great antiquity".

I don't accept that the last
comment automatically follows the first. To me it looks as though R1b went west from Central Asia, Q went east from Central Asia and R1a stayed put, although R1a had already coalesced in India by that time.

"Ancient hgs such as C and D-E typically show a phenomenal geographic spread".

Agreed. And I'd guess that Y-hap Q first shot past Y-hap C3 on its way to America. Then Y-hap C3 followed on.

eurologist said...

Top five reasons why Anatolia is not the homeland of proto-IE:

1. If it were, you would expect a spread from there North, East, or South. Nothing of that sort happened for at least ~4,000 years.

2. If it were, the immediate western impact would have been a (proto-) IE Greece, followed by a (proto-) IE spread into Italy and Spain, along the well-known Mediterranean path of the advance of agriculture. We know this did not happen - all of these countries by pretty much all accounts did not become IE until at least ~4,000 years later.

3. All quantitative analyses put the "origin" of proto-IE later than the development of agriculture in Anatolia (even if as little as 1,000 - 2,000 years). Note, however, that dating around the origin can be misleading if the initial group was limited in size and localized and did not spread for a long interval. That is, the origin of proto-IE can go back millennia before we define what characterizes it, and when its children spread.

4. If you accept from the data and datings that the Kurgan hypothesis can only apply late and to the East (Slavo-Baltic and Indo-Iranian), while these branches always appear downstream in quantitative measurements, you also must accept that there is a much earlier earlier source to the West - but one that is by no known route and timing connected to the South (Anatolia).

5. The single most well-documented early neolithic European population explosion was LBK. Clearly, anything after was just a modulation of theme, and did not reach this wide expanse from the Seine to the Ukraine. Occam's razor waves its hand.

Dienekes said...

1. If it were, you would expect a spread from there North, East, or South. Nothing of that sort happened for at least ~4,000 years.

Not sure what empirical data on this you have. It's like saying IE speakers did not spread from Anatolia because they didn't spread from Anatolia/


2. If it were, the immediate western impact would have been a (proto-) IE Greece, followed by a (proto-) IE spread into Italy and Spain, along the well-known Mediterranean path of the advance of agriculture. We know this did not happen - all of these countries by pretty much all accounts did not become IE until at least ~4,000 years later.

Maritime pioneer colonization is not expected to have quite the same effect. We must also avoid the temptation of assuming that only IE groups were involved in all farming dispersals.


3. All quantitative analyses put the "origin" of proto-IE later than the development of agriculture in Anatolia (even if as little as 1,000 - 2,000 years). Note, however, that dating around the origin can be misleading if the initial group was limited in size and localized and did not spread for a long interval. That is, the origin of proto-IE can go back millennia before we define what characterizes it, and when its children spread.

Methods date the first split within the language family, that is what is meant by "origin". The first split is not expected to be contemporaneous with the invention of agriculture, but with the first separation of speakers of the language, and the dates are quite consistent with the Neolithic colonization of Europe.

4. If you accept from the data and datings that the Kurgan hypothesis can only apply late and to the East (Slavo-Baltic and Indo-Iranian), while these branches always appear downstream in quantitative measurements, you also must accept that there is a much earlier earlier source to the West - but one that is by no known route and timing connected to the South (Anatolia).

Not sure what you are saying here. I don't see any reason to associate the Kurgan phenomenon with either Balto-Slavic or Indo-Iranian.

What seems to have happened is that the steppe populations, whose early language is unknown, seem to have been Iranized, at least partially, by the 1st millennium BC. 1-2 thousand years later they were Turkicized. If the Soviet Empire had not collapsed, they'd been Russianized probably. It seems that the steppe populations are quick to grab a new language, which can then spread easily among them.

5. The single most well-documented early neolithic European population explosion was LBK. Clearly, anything after was just a modulation of theme, and did not reach this wide expanse from the Seine to the Ukraine. Occam's razor waves its hand.

Not sure how you are invoking Occam's razor here. The LBK was probably IE-speaking, although we can't know for sure what languages they spoke. It's possible that later population movements in North-Central Europe established the dominant languages spoken there today. After all, Slavic grew to a huge size from an origin that wasn't even noticed by classical historians, so it's quite possible that there were episodes of language replacement after the LBK that we simply don't know about.

German Dziebel said...

@Terry

"And I don't accept for a moment that they entered America together."

All genetic publications claim that hg C and Q came to the New World as a single wave. And I can see one of their points: hg C is widely distributed in the New World, albeit at low frequencies. It's very easy to imagine that hg Q simply began to recently dominate what once was a more evenly distributed genetic landscape.

"Q went east from Central Asia"

Q is also found in West Asia and northern Europe.

"That explains the gap between Q in the Ket/Selkup people and the Americans."

Exactly. This is the pre-NO stratum.

"And I'd guess that Y-hap Q first shot past Y-hap C3 on its way to America. Then Y-hap C3 followed on."

I can't accept this fantastical scenario.

ashraf said...

"then this language family would be old enough to transcend some of the racial divides."

The Nostraticists I read their books(bomhard+bernal+renfrw) connect the developpement of nostratic out of "out of african" with the natufian bioculture of around 15.000 bc and I think the nostratic tongue is connected with the caucasoid type (especially the west asian component) the affinity of afrasan-kartvelian-IE with altaic-uralosiberian-eskimoaleut is the result of linguistic influences from the expanding indo and para-indohitties upon hunter gatherer societies of Siberia
Here below a link for a section taken from Bomhard's book.(please read the quoted section)

http://s1.zetaboards.com/anthroscape/single/?p=552232&t=3812504

ashraf said...

"The basic pattern of the expansion of the Homo sapiens sapiens on the Eurasian continent, of “the peopling of Eurasia,” is of the “ripple” type, with the Near East in the center pumping waves after waves of immigrants to the east (East Asia), west (Europe), and somewhat later, namely after the retreat of the last ice-age, north and south (back to North Africa). The model of Eurasia-peopling is, that is to say, “Out of the Near East Again and Again.” The first wave has just been described and the linguistic vestiges of it identified. Around 40,000 years or so BP, a second wave of immigration

We now come to the “third wave,” the third “Out of Near East.” Most of today’s linguistic and ethnic diversity on Eurasia continent in fact had its origin in this third, Nostratic focal expansion. Let me read to you from the excellent statement on this topic, Allan R. Bomhard and John C. Kern’s The Nostratic Macrofamily (Mouton de Gruyter, 1994). First, his concluding statement on the previous, second focal expansion, that of Dene-Caucasian.
“As always in hunter-gatherer societies, mobility was at a premium. Canoes were used for water travel and snow shoes and sleds were developed for overland travel in winter. The conditions were favorable for the rapid spread of tribes and their new linguistic family over immense distances. This expansion, which is called Mesolithic, is indicated archaeologically by microliths found all along Northern Eurasia and Southward through the Caucasus into the Near East, where it later developed smoothly into the Neolithic with its domestication of cereals and of animals suitable for food and fibers.
“The Mesolithic culture is aptly named, for it provided a gradual though rapid transition between the Upper Paleolithic and the agricultural Neolithic. There was, in fact, a steady advance in man’s ability to control and exploit his environment…”
the Mesolithic culture, with its Nostratic language, had its beginning in or near the Fertile Crescent just south of the Caucasus, with a slightly later northern extension into Southern Russia in intimate association with woods and fresh water in lakes and rivers. From these positions, it had ready access to the lower Danube and the Balkans (Indo-European), to the Caucasus (Kartvelian), south of the Caucasus into Mesopotamia, Palestine, Egypt, and the rest of North Africa (Sumerian and Afroasiatic), eastward into Central Siberia (Elamo-Dravidian), and northward and thence eastward along the Circumpolar fringe (Uralic-Yukaghir, Altaic, Chukchi-Kamchatkin, Gilyak, and Eskimo-Aleut [: these together with Indo-European constitute the Eurasiatic subgroup within Nostratic). In the process of its expansion, it undoubtedly effected a linguistic conversion of many tribes of Dene-Caucasian or other origins; this accounts for the fact that non-Nostratic languages in Eurasia in historic times have been found mostly as relics in mountainous regions. Exceptions are Chinese and the now moribund or extinct Ket, which, together with Hattic and Hurrian, probably represent post-Nostratic reemergences of Dene-Caucasian speakers from their relic areas"

German Dziebel said...

@ashraf

"I think the nostratic tongue is connected with the caucasoid type (especially the west asian component) the affinity of afrasan-kartvelian-IE with altaic-uralosiberian-eskimoaleut is the result of linguistic influences from the expanding indo and para-indohitties upon hunter gatherer societies of Siberia"

I think what the Nostraticists you quoted mean is that Nostratic languages spoken by a generalized Caucasoids expanded from an area south of the Caucasus into Sub-Saharan Africa, Europe and Asia. As they expanded, they absorbed native elements on all three continents. That's why Sub-Saharan Afroasiatic speakers are black, Eskimos and Altaics are Mongoloid, Dravidians are South Asians in type, etc. So, the languages diverged through the natural process of linguistic differentiation from a single source, whereas the phenotypes on the fringes of the Nostratic expansion come from indigenous people. So, the Nostratic language possesses integrity throughout its range of distribution. This is very different from your earlier claim that Uralic and Altaic languages derive their pronouns and verbal endings from a Sprachbund.

ashraf said...

I think by the time of nostratic expansion the ancestor of proto uralo-siberian and proto altaic speakers were inhabiting eastern areas of northern asia and did not come in interaction with the nostratic peoples, otherwise we ,nowadays, would have seen a caucasoid inpuut widespread on all altaic and uralo-siberian (and eskimo-aleut) speaking populations, probably it's when those populations begin to migrate westward into central asia , that they did take loanwords+pronouns from a nostratic/para IE and this is shown by the fact that while the altaic and uralic folks that expanded westward do have some caucasoid input, the ones that remained in the altaic homeland (manchuria) turkic homeland (lena river) and uralic homeland (ob river) are lacking caucasoid input.
However in eastern and northern africa all afrasan speakers do sahre from 20 to 40% west asian input (somalis and ethiopians included)
We know that there is an important iranian influence on uralic and it could be very well that an earlier para IE or nostratic branch did provide uralic and altaic with pronouns[especially when they do have a nostratic IE etymology as "man" (I) to be connected with "mannus" (man)]

ashraf said...

from wikipedia about the possible nostratic homeland

"Allan Bomhard and Colin Renfrew are in broad agreement with the earlier conclusions of Illich-Svitych and Dolgopolsky in seeking the Nostratic Urheimat (original homeland) within the Mesolithic (or Epipaleolithic) in the Fertile Crescent, the stage which directly preceded the Neolithic and was transitional to it.

Looking at the cultural assemblages of this period, two sequences in particular stand out as possible archeological correlates of the earliest Nostratians or their immediate precursors. Both hypotheses place Proto-Nostratic within the Fertile Crescent at around the end of the last glacial period.

The first of these is focused on the Levant. The Kebaran culture (18,000–10,500 BCE) not only introduced the microlithic assemblage into the region, it also has African affinity, specifically with the Ouchtata retouch technique associated with the microlithic Halfan culture of Egypt (24,000–17,000 BCE).[23] The Kebarans in their turn were directly ancestral to the succeeding Natufian culture (10,500–8500 BCE), which has enormous significance for prehistorians as the clearest evidence of hunters and gatherers in actual transition to Neolithic food production. Both cultures extended their influence outside the region into southern Anatolia. For example, in Cilicia the Belbaşı culture (13,000–10,000 BCE) shows Kebaran influence, while the Beldibi culture (10,000–8500 BCE) shows clear Natufian influence.
The second possibility as a culture associated with the Nostratic family is the Zarzian (12,400–8500 BCE) culture of the Zagros mountains, stretching northwards into Kobistan in the Caucasus and eastwards into Iran. In western Iran, the M’lefatian culture (10,500–9000 BCE) was ancestral to the assemblages of Ali Tappah (9000–5000 BCE) and Jeitun (6000–4000 BCE). Still further east, the Hissar culture has been seen as the Mesolithic precursor to the Keltiminar culture (5500–3500 BCE) of the Kyrgyz steppe.
It has been proposed that the broad spectrum revolution[24] of Kent Flannery (1969),[25] associated with microliths, the use of the bow and arrow, and the domestication of the dog, all of which are associated with these cultures, may have been the cultural "motor" that led to their expansion. Certainly cultures which appeared at Franchthi Cave in the Aegean and Lepenski Vir in the Balkans, and the Murzak-Koba (9100–8000 BCE) and Grebenki (8500–7000 BCE) cultures of the Ukrainian steppe, all displayed these adaptations.

Bomhard (2008) suggests a differentiation of Proto-Nostratic by 8,000 BCE, the beginning of the Neolithic Revolution in the Levant, over a territory spanning the entire Fertile Crescent and beyond into the Caucasus (Proto-Kartvelian), Egypt and along the Red Sea to the Horn of Africa (Proto-Afroasiatic), the Iranian Plateau (Proto-Elamo-Dravidian) and into Central Asia (Proto-Eurasiatic, to be further subdivided by 5,000 BCE into Proto-Indo-European, Proto-Uralic and Proto-Altaic)."

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

German Dziebel said...

@ashraf

"I think by the time of nostratic expansion the ancestor of proto uralo-siberian and proto altaic speakers were inhabiting eastern areas of northern asia and did not come in interaction with the nostratic peoples, otherwise we ,nowadays, would have seen a caucasoid inpuut widespread on all altaic and uralo-siberian (and eskimo-aleut) speaking populations, probably it's when those populations begin to migrate westward into central asia , that they did take loanwords+pronouns from a nostratic/para IE and this is shown by the fact that while the altaic and uralic folks that expanded westward do have some caucasoid input, the ones that remained in the altaic homeland (manchuria) turkic homeland (lena river) and uralic homeland (ob river) are lacking caucasoid input."

This is a good point, although Ugric populations have a variable "European" and "Asian" components in both mtDNA and Y-DNA. Also, Uralic and Mongoloid physical types seem to be less than 10,000 years ago, which is well within the range of the hypothetical Nostratic timeline. And once again, your idea of pronoun and verb ending borrowing is interesting but it contradicts Nostratic theory.

"Allan Bomhard and Colin Renfrew are in broad agreement with the earlier conclusions of Illich-Svitych and Dolgopolsky in seeking the Nostratic Urheimat (original homeland) within the Mesolithic (or Epipaleolithic) in the Fertile Crescent, the stage which directly preceded the Neolithic and was transitional to it."

Thanks for the quote from Wiki. I got a god picture of the way Nostraticists see the world. I know Allan Bomhard, and I even arranged a publication for him in a journal devoted to kinship studies. See here we're part of the same volume (http://www.libex.ru/detail/book444543.html). But I wasn't familiar with the details of how Nostraticists map their hypothesis onto archaeology. BTW, what's your take on high frequency of such a "Mongoloid/Amerindian" marker as shavel shaped incisors in Catalhouyuk (etd.ohiolink.edu/send-pdf.cgi/Pilloud%20Marin%20Anna.pdf?...dl=y)? Shovel shaped incisors are also frequent in India but the they slowly peter away as one moves into Europe and Africa? I wondered if this signifies a major expansion from east to west.

terryt said...

"It seems that the steppe populations are quick to grab a new language, which can then spread easily among them".

Of course that means that people through the region may have spoken early dialects of Proto-Indo-European at some early period. I tend to be in agreement with Eurologist here.

"hg C is widely distributed in the New World, albeit at low frequencies".

But it's far more common in the north, unlike Q which is spread completely through both North and South America.

"It's very easy to imagine that hg Q simply began to recently dominate what once was a more evenly distributed genetic landscape".

To me C has all the appearance of being a later arrival than Q.

"Q is also found in West Asia and northern Europe".

So presumably it also went west from Central Asia once its ancestor P had reached that region. As did its brother haplogroup R1b.

"Exactly. This is the pre-NO stratum".

We agree on this then. C3 and Q passed to the north of NO, then N and some O moved north. We just disagree as to whether C3 and Q moved east together or separately.

"I can't accept this fantastical scenario".

I can't see at all why you would call it a 'fantastical scenario'. To me it's the only solution that makes sense.

Waggg said...

Ashraf : "the ones that remained in the altaic homeland (manchuria) turkic homeland (lena river) and uralic homeland (ob river) are lacking caucasoid input."

For the records, some caucasoid input is found in north Siberia up to the Yakuts! (the latter have some Y-DNA R1a * they also have 8.4% haplotypes Caucasoid mtDNA haplogroups ** (H, HV1, J, T, W and U (I remember there was a U51b1b for instance)) (they also have a little north European component in their autosomal profile). Puzyrev et al 2003 had found 6% of caucasoid mtDNA hgs.

* http://www.ncbi.nlm.nih.gov/pubmed/12942782

** http://evolutsioon.ut.ee/publications/Fedorova2003.pdf

And of course the Uralic/Ugric population east of the Ural mountains have got some Caucasoid input as well.
example : http://www.ncbi.nlm.nih.gov/pmc/articles/PMC379094/

I'm not saying it has anything to do with the raised linguistic considerations though.

German Dziebel said...

@terry

"To me C has all the appearance of being a later arrival than Q."

This used to be the case with mtDNA hg B. Originally it also had "all the appearance of a late arrival." Coastal arrival in its case. But all the publications from 1995 on pretty much consider it part of the original Amerindian gene pool. Appearances are deceptive, Terry, as they are not based on anything.
The spread of hg C is broad enough (Na-Dene, Eskimo, Southeast, South America) to treat it's antiquity in the Americas as a null hypothesis. Compare hg B is very infrequent in the North - but not because it was never there but because it was either replaced by later haplogroups or drifted out in small populations. Y-DNA hg Q is very rare in Siberia but very frequent in the Americas; hg C is frequent in Siberia and rare in the Americas. It's just an artifact of demography and not a straightforward indication of age.

mathilda said...

I'm going with Anatolian Neolithic PIE as the primary source (9kbp), with a later dialet north of the black sea (5-6kbp) giving rise to the Tocharian and later Balkan dialect branches.

Also places Anatolian Neolithic PIE close enough to share words with very early Afro Asiatic languages like Proto Semitic. Other later/northerly scenarios can't do that.

George said...

G2a3 6000 years ago was in Germany.
I think that
The G2a3b1a2 is the European-IE haplogroup.
The G2a3b(XG2a3b1a2) are the "Indo-Albanians".
The G2a3a are the "Greko-Armenians".
G2a3*(XG2a3a,G2a3b) are the pre-Tocharians.
And G2a(XG2a3) are the Hittites.

http://3.bp.blogspot.com/-UK6uv6RZBlw/TZr4U1kHkKI/AAAAAAAADbY/jvqITDejMG0/s1600/nature.jpg

Unknown said...

It's interesting to see how David Anthony responded to this study in the times quote below ("I see the wheeled-vehicle evidence as a trump card over any evolutionary tree..") Because the wheel arrives so late in the archaeology -- compared to the domesticated horses he used to endorse -- this sets a later date for a unified IE language than Steppes advocates like Mallory have recently been taking -- usually about 4500BC.

I'd agree that it makes no sense to think these "wheeled vehicle" words were all neologisms that had no pre-existing meaning in IE. And therefore they don't have much use in matching linguistic to the archaeological data. A pertinent example are the earliest potter wheels, tournettes, were already in use in the Near East around 4500BC

What does surprise me a bit is how invested the Steppes theory is. It's a little ironic that V. Gordon Childe, who first originated the theory, ended up rejecting it and favoring Anatolia as well.


"A key piece of their evidence is that proto-Indo-European had a vocabulary for chariots and wagons that included words for “wheel,” “axle,” “harness-pole” and “to go or convey in a vehicle.” These words have numerous descendants in the Indo-European daughter languages. So Indo-European itself cannot have fragmented into those daughter languages, historical linguists argue, before the invention of chariots and wagons, the earliest known examples of which date to 3500 B.C...
'I see the wheeled-vehicle evidence as a trump card over any evolutionary tree,' said David Anthony, an archaeologist at Hartwick College who studies Indo-European origins."
www.nytimes.com/2012/08/24/science/indo-european-languages-originated-in-anatolia-analysis-suggests.html