Revising the classification of Indo-European languages

Maciamo

Veteran member
Admin
Messages
9,970
Reaction score
3,273
Points
113
Location
Lothier
Ethnic group
Italo-celto-germanic
We have hypothesised in Germanic words of non-IE origin that Proto-Germanic borrowed a few common words from indigenous pre-IE Scandinavians. I believe that there may be a much bigger proportion of Latin and Greek words (including those inherited in modern Romance languages) that are not Indo-European at all, but came from Near Eastern languages. This wouldn't be surprising considering the very substantial percentage of Near-Eastern DNA in Italy and Greece. I am often perplexed to see how linguists assume that if words in Germanic or Slavic languages are similar to those in Latin and Greek then they must be of IE origin. This shouldn't necessarily be the case, as I will explain.

For a start, I am inclined to think that the fusion between two populations speaking different languages will result either in :

1) the thorough hybridisation (creolisation) of the two languages, as was the case with Middle English and Norman French (fusing to become modern English), or with Jomon Japanese and Yayoi Sino-Korean (merging into one another to make modern Japanese). I think this is what happened with Greek and Latin, probably hybridising PIE with a West Asian language (possibly associated with Bronze Age J2 people).

2) one language becomes dominant but absorbs some vocabulary from the other language. The dominant language can be the one of the newcomer (e.g. English almost completed erased Gaelic in Scotland, but Scottish English has kept plenty of loanwords from Gaelic) or the indigenous one (e.g. the Franks eventually adopted Latin, but modern French has hundreds if not thousands of words of Frankish origin). I think that this is what happened with Proto-Celtic and Proto-Germanic (dominant IE with indigenous loanwords).

The third option would be that one language (presumably the one of the newcomers) completely wipes all traces of the other languages. But I do not think that this is possible unless one population (presumably the indigenous) is totally exterminated. But has it ever happened ? Europeans can close in the North America and Australia, but despite the huge disparity in technological advancement American and Australian English borrowed hundreds of words from Amerindians and dozens from aboriginal Australian.

Surely, if the ancestry of the ancient Greeks and Romans is over 50% non-Indo-European, as now transpires from the insight provided by population genetics, a very sizeable part of their language must also be of non-IE origin.

I also believe that the way linguists have reconstructed Proto-Indo-European language is faulty because they assumed that languages fit only in one family. So much is obvious when so many linguists stubbornly refuse to classify English as both Germanic and Romance, and insist it is only Germanic when 70% of its vocabulary is Romance (including Greek loanwords in Latin and French) and its grammar is now closer to Romance than Germanic languages.

I also believe that there wasn't one PIE language, but two. There were clearly two separate ethnic groups that spoke and diffused Indo-European languages: one originated in Anatolia and around the Black Sea under the dominant male lineage R1b, and one originating in the forest-steppe of Ukraine, Belarus and Russia led by R1a men. The two groups converged into one political and/or cultural entity, probably during the Yamna period, but did not mix much genetically. There must have been an exchange of ideas, technology, religion and vocabulary, but the two groups remained distinct and evolved separately, each migrating to different regions (at first, Southeast, Central and Western Europe for R1b, and Northern Europe, Central and South Asia for R1a).

As the two groups did not share a common ancestor since about 25,000 years ago, they could not have spoken the same language, nor anything remotely similar. They first "met" each others when R1b people moved from Anatolia to the Pontic Steppes, perhaps 6000 years ago. Since only a small minority of of R1b people joined the ranks of R1a population, and even less (if any) R1a moved with the R1b population in the Pontic Steppes during the Yamna period, logic dictates that the two groups spoke different languages, although probably with a significant amount of loan mutual words (probably with a bias towards more R1b words assimilated into the R1a language).

Linguists have classified IE languages well before the recent discoveries of population genetics. Yet, the branches associated with haplogroup R1a, Baltic, Slavic, Indo-Aryan and Iranian, were all classified together under the Satem branch. This is not a coincidence. Satem languages really are quite different from Centum languages, not just because of a shift of pronunciation, but also because a lot of words common to Satem languages have no cognate in Centum languages (and vice versa). If it hasn't been done, it would be interesting to calculate the percentage of words that are undeniably (and I insist on that) related to each others within Satem, within Centum, and across Satem and Centum. Those that are only vaguely similar should be eliminated. I am confident that this will create a proper rift between the two groups.

Obviously some intermediary languages like Greek and Armenian should be listed separately because they are 1) genetically mixed R1a-R1b, 2) too close to both R1a and R1b homelands, and 3) strongly influenced by other West Asian languages. Germanic languages will also show similarities between both Satem and Centum groups, and indeed may be seen as the perfect link between the two groups based on haplogroup distribution. Proto-Celtic should be seen as the epitome of the R1b-Centum branch, while Latin should be approached cautiously as a hybrid language whose vocabulary may be unrepresentative of IE languages.

There is also the possibility of vocabulary surviving from Neolithic Near-Eastern languages. If Basque is a direct descendant of a Neolithic (or even Mesolthic/Paleolithic) language, then there is no reason why plenty of Neolithic loanwords shouldn't have survived all over Europe. These can be easily misinterpreted for Indo-European words due to their pan-European distribution. We nearly don't know anything about Neolithic languages (as writing didn't exist), but they shouldn't be discarded too easily. After all, Indo-European languages spread before the time of writing too.

In the Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World by J.P. Mallory and D.Q. Adams I was shocked to see how many common words have many completely different PIE roots listed. For example, the PIE words for goat are *bhugos, *diks, *heli, *h(a)egos, *h(a)egs, *kapros, *(s)kegos, *ghaidos and *kogheh. For grain we find linguistic abominations like *drh(x)weh(a)-, *dhoh(x)neh(a)-, *grh(a)nom, h2/3(e)lg(h)-, *meig(h)-, *proksom, *ses(y)o-, *yew(e)s, *pitus, *h2ed-, *melh2-. And I can't encode all the accents and circles that come with that !

What we should ask ourselves is why would a rather primitive Bronze Age society have ten words to designate something as simple as a goat or grain ? The best explanation is that these words come from various Neolithic languages assimilated by the Indo-Europeans. I noticed that words that have the most roots listed are typically those associated with the Neolithic lifestyle (furrow, rain, dry, wet, grain, fruit, pot, milk, goat, animal skin) and very basic verbs (burn, cut, move, jump, run, turn, pull, push, shake, speak, cry). Besides many of these reconstruction are unpronounceable because they try to find a common root for words for certainly don't have one.

There three PIE words for cow listed in the above-mentioned book : *g(w)ous (Lat bos, NE cow, Grk bous, Skt gau), h1egh- (Skt ahi) and *wokeh(a)- (Lat vacca, Skt vasa). I am almost sure that the last one is not Indo-European, but rather related to the Mesopotamian language spoken the Neolithic or Bronze Age people associated with Y-haplogroup J2. It cognates with the Biblical Hebrew baqar and the modern Arabic baqara. I suggested earlier that the J2 people were associated with the diffusion of cattle and bull worship. J2 expanded from West Asia all the way from Italy to India, so it is far more likely that the Latin vacca and the Sanskrit vasa derive from this West Asian language of the J2 folk. The Sanskrit ahi is probably a native term, completely unrelated to PIE.

I know that it is a lot of unconventional (some will say heretic) ideas in one time. But I really want to test this hypothesis because nobody has reviewed the Indo-European language tree since we know about the genetic R1a-R1b division.

I will start a thread about Latin & Greek words of non-Indo-European origin where I invite you to add your suggestions.
 
Last edited:
my only argue is the Greco-Aryan

according that Greeks and Medes were from the same linguistic origin
except sanshqrit all these languages share a unigue grammar and forms (voices, numbers, cases)
yet modern Greek from Homer's time are Centum while west iranian are satem,
but they have kept the grammar, until 800 BC
on the other hand we find many IE languages that do not even have full virb's future time,

although I agree that modern IE is a language that assimilated earlier languages, though it changed sounds according the sub-stractum of earlier language and the need of communications,

the case of Greco-aryan is clear a prove of 1 linguistic Group exist and not 2.

although it will be interesting,
 
Maciamo, I have criticism of the following issues:

Linguists have classified IE languages well before the recent discoveries of population genetics. Yet, the branches associated with haplogroup R1a, Baltic, Slavic, Indo-Aryan and Iranian, were all classified together under the Satem branch. This is not a coincidence. Satem languages really are quite different from Centum languages, not just because of a shift of pronunciation, but also because a lot of words common to Satem languages have no cognate in Centum languages (and vice versa). If it hasn't been done, it would be interesting to calculate the percentage of words that are undeniably (and I insist on that) related to each others within Satem, within Centum, and across Satem and Centum. Those that are only vaguely similar should be eliminated. I am confident that this will create a proper rift between the two groups.

The Centum/Satem split is a bit more complicated than that, and you have to consider what it really means. First off, in addition to Armenian (which you address below), you also forgot to mention Albanian which is also a Satem language. It has to be added, and I elaborated on that in the Albanian thread, the non-native vocabulary in Albanian is substantial, most importantly from Latin. You also have to consider the existence of Dacian and Thracian.

Then, you have to consider what the shift is all about, namely the treatment of a set of three sounds found in PIE called the palatovelars: *k´, *g´, *g´h

In the Centum languages (Celtic, Germanic, Greek, Italic, Tocharian), these are merged with the plain velars *k, *g, *gh.

In the Satem languages (Albanian, Armenian, Balto-Slavic, Indo-Iranic), these are turned into fricative sounds.

Then there is the timing of the Centum-Satem split to consider. The earliest we have the Centum/Satem split attested from the mid-2nd millennium BC (Mycenean Greek, representative for the Centum languages, and the "Mitanni" Indo-Iranic loanwords into Hurrian as a representative of Satem. We also must consider that the texts of Rigvedic Sanskrit, even though they are attested much later, must actually date from about this time period.

The question is, how old can the split be? We have common words for 'horse' (*ek´wos), 'hundred' (*k´mtom), but we do have a common word for "gold" in Germanic and Balto-Slavic (English "gold" vs. Latvian "zelta" vs. Russian "zoloto"), and it clearly predates the Centum-Satem split. My opinion is that the Centum-Satem split cannot have occured before the mid-3rd millennium BC.

I also do not see the point of your insistence of "undeniability". You have to consider that new words can be (and indeed, have been frequently) formed inside distinct IE branches from previous common words, but that acquired new meanings in the individual branches. If we discard these words, we give ourselves onto a very slippery path.

Obviously some intermediary languages like Greek and Armenian should be listed separately because they are 1) genetically mixed R1a-R1b, 2) too close to both R1a and R1b homelands, and 3) strongly influenced by other West Asian languages. Germanic languages will also show similarities between both Satem and Centum groups, and indeed may be seen as the perfect link between the two groups based on haplogroup distribution. Proto-Celtic should be seen as the epitome of the R1b-Centum branch, while Latin should be approached cautiously as a hybrid language whose vocabulary may be unrepresentative of IE languages.

Honestly, I don't see Latin as that "unrepresentative" at all, especially if you consider the similarities with Celtic on the one hand, and with Greek on the other.

There is also the possibility of vocabulary surviving from Neolithic Near-Eastern languages. If Basque is a direct descendant of a Neolithic (or even Mesolthic/Paleolithic) language, then there is no reason why plenty of Neolithic loanwords shouldn't have survived all over Europe. These can be easily misinterpreted for Indo-European words due to their pan-European distribution. We nearly don't know anything about Neolithic languages (as writing didn't exist), but they shouldn't be discarded too easily. After all, Indo-European languages spread before the time of writing too.

As I have said before, if we substract all the IE-borrowed words from Basque (principally Latin and Romance, but also a small share of Celtic words) we end up with a language that seemingly cannot be older than the Chalcolithic, and since it is an isolate language nowadays we have no way to determine the nature of these "native" words. I'd also like to point out that the diversity of non-IE languages in Antiquity in Iberia suggests that the diversity of pre-IE languages in Europe may have been rather large, and that Basque may not be representative at all of the languages that were spoken elsewhere.

In general, while I absolutely agree that loanwords from elsewhere exist, and that they probably follow patterns, I must absolutely disagree on the idea of "two" Indo-European languages. Indo-Europeanists reconstruct one proto-language, and this is not without a reason.

I also think that you are definitely overinterpreting the genetic evidence here (because I am reasonably certain you cannot press the R1b/R1a division onto the Indo-European languages), but I would like to say that this is my opinion, and I'm definitely willing to offer my advice here. :)
 
The Centum/Satem split is a bit more complicated than that, and you have to consider what it really means. First off, in addition to Armenian (which you address below), you also forgot to mention Albanian which is also a Satem language. It has to be added, and I elaborated on that in the Albanian thread, the non-native vocabulary in Albanian is substantial, most importantly from Latin. You also have to consider the existence of Dacian and Thracian.

I know. I wasn't going to list all IE languages. I was trying to make a point and it is unnecessary to review each language one by one for that. Obviously Albanian has a lot of non-IE loanwords, perhaps the most of any IE language.

Then, you have to consider what the shift is all about, namely the treatment of a set of three sounds found in PIE called the palatovelars: *k´, *g´, *g´h

In the Centum languages (Celtic, Germanic, Greek, Italic, Tocharian), these are merged with the plain velars *k, *g, *gh.

In the Satem languages (Albanian, Armenian, Balto-Slavic, Indo-Iranic), these are turned into fricative sounds.

Then there is the timing of the Centum-Satem split to consider. The earliest we have the Centum/Satem split attested from the mid-2nd millennium BC (Mycenean Greek, representative for the Centum languages, and the "Mitanni" Indo-Iranic loanwords into Hurrian as a representative of Satem. We also must consider that the texts of Rigvedic Sanskrit, even though they are attested much later, must actually date from about this time period.

The question is, how old can the split be? We have common words for 'horse' (*ek´wos), 'hundred' (*k´mtom), but we do have a common word for "gold" in Germanic and Balto-Slavic (English "gold" vs. Latvian "zelta" vs. Russian "zoloto"), and it clearly predates the Centum-Satem split. My opinion is that the Centum-Satem split cannot have occured before the mid-3rd millennium BC.

I know all this. I have read the Oxford Introduction to Proto-Indo-European as well as most of your posts on the forum. I think that you fail to see what I meant. It doesn't matter how the satemisation of IE pronunciation happened, nor when it happened. The new idea here is that R1a people (represented chiefly by the Balto-Slavic and Indo-Iranian branches, as other branches also have sizeable R1b populations) could not have spoken the same language as R1b people.

I also do not see the point of your insistence of "undeniability". You have to consider that new words can be (and indeed, have been frequently) formed inside distinct IE branches from previous common words, but that acquired new meanings in the individual branches. If we discard these words, we give ourselves onto a very slippery path.

Isolated words in one language or one family (e.g. Western Romance) could indeed be new words coined after the expansion of IE languages. But then they shouldn't be considered Proto-Indo-European, or even Indo-European. They are just local dialects.

Isolated words could just as well be older loanwords from pre-IE languages. There is no way to know if it is a new coinage or an ancient borrowing.

If we do find similar words that clearly aren't pan-Indo-European and that these words are distributed in several languages in a same region sharing some common non-IE ancestry, then one could wonder whether this word just doesn't come from another (pre-IE) language. This is the point I was trying to make with the Latin vacca, with relatives in other regions with high frequencies of haplogroup J2, both in Indo-European and Semitic languages (that doesn't mean that the word is Semitic if it isn't IE; in this case it is probably from a completely unrelated language that was replaced by Semitic and IE languages).

I wouldn't consider that a word that is completely missing from a main branch (e.g. Italo-Celtic or Indo-Iranian) of the IE tree is truly PIE, but more probably borrowed from another language.

Honestly, I don't see Latin as that "unrepresentative" at all, especially if you consider the similarities with Celtic on the one hand, and with Greek on the other.

Greek is just as unrepresentative as Latin, if not more. How could Greek language be overwhelmingly Indo-European in origin if only 25% of male lineages in Greece are Indo-European and an even smaller percentage of maternal lineages ? Most Greek R1a being close to the Slavic one, early Middle Ages, and which therefore didn't influence ancient Greek Besides, Indo-European speakers conquered ancient Greece at least twice (Mycenaeans and Dorians), as reflected by the shift from Q-Greek to P-Greek that you explained to me.

It's one thing to hypothesise how languages spread 5000 or 6000 years ago. But it's always good to compare what really happened in historical cases. I do not know any proven historical example of a small minority of people that successfully imposed their language on conquered people and eradicated the native language. There are at least three major obstacles to this in the case of ancient Greece :

1) The Indo-European conquerors that first set foot in Greece, the Mycenaeans, probably didn't make up more than 5% to 10% of the population after the conquest. They were a small ruling elite. So were the Dorians later. They could have imposed their language for administration, but a major native substrate must have survived.

2) the conquered people (pre-IE Greeks) belonged to a quite advanced civilisation for the time. The Indo-Europeans were only superior militarily. Proud civilizations don't give up their language easily to "barbarians".

3) The Indo-Europeans brought a completely new and unrelated language to Greece, which must have been difficult for the natives to learn, even if there was a political will to do so. This argument is well demonstrated by Nicholas Ostler in Empire of the Words. The Romans never succeeded in imposing Latin to the common people in the eastern half of their empire because a) Greek speakers felt too proud/superior to give up their language, and b) Latin was too different and difficult to learn for the uneducated masses of Semitic speakers in the Near East. The Arabs were the only ones who managed to replace local languages in the Levant, Mesopotamia and North Africa because Arabic was sufficiently close to Aramanic, Egyptian, Berber, etc. to be learnt easily by everyone.

My point here is that Greece has a too small percentage of Indo-European ancestry for Greek to be almost purely Indo-European in origins. It's hard to see how more than half of Greek words could be of Indo-European origin. Latin might have a bit more, but it depends if we take Classical Latin or Vulgar Latin.


As I have said before, if we substract all the IE-borrowed words from Basque (principally Latin and Romance, but also a small share of Celtic words) we end up with a language that seemingly cannot be older than the Chalcolithic, and since it is an isolate language nowadays we have no way to determine the nature of these "native" words. I'd also like to point out that the diversity of non-IE languages in Antiquity in Iberia suggests that the diversity of pre-IE languages in Europe may have been rather large, and that Basque may not be representative at all of the languages that were spoken elsewhere.

I didn't suggest in any way that Italic, Celtic or Germanic languages borrowed words from the ancestral language of Basque. I said from Neolithic languages, of which there were plenty, and many were unrelated or only very distantly related. When we see the number of non-IE West Asian languages that existed until the late Bronze Age (Hurrian, Urartian, Hattian, Kassite, Sumerian, Elamite...), it wouldn't surprising if Neolithic Europe and West Asia had hundreds of languages.

In general, while I absolutely agree that loanwords from elsewhere exist, and that they probably follow patterns, I must absolutely disagree on the idea of "two" Indo-European languages. Indo-Europeanists reconstruct one proto-language, and this is not without a reason.

I also think that you are definitely overinterpreting the genetic evidence here (because I am reasonably certain you cannot press the R1b/R1a division onto the Indo-European languages), but I would like to say that this is my opinion, and I'm definitely willing to offer my advice here. :)

Well, everything you have learnt about Indo-European linguistics was written by people who had zero knowledge of current population genetics. If they build their whole theory on mistaken foundations, there will be a lot to reconsider.
 
I know. I wasn't going to list all IE languages. I was trying to make a point and it is unnecessary to review each language one by one for that. Obviously Albanian has a lot of non-IE loanwords, perhaps the most of any IE language.

Yeah, I see. I think it makes sense to treat the most well-established groupings (that is, Balto-Slavic and Indo-Iranic) as single entities in this discussion, however.

The funny thing is, I don't actually think that Albanian has that many non-IE words. Most of the loanwords in Albanian are from other IE languages (in chronological order: Classical Greek, Latin, East Germanic, Slavic), and from what I have seen, most non-IE words in Albanian are relatively recent additions dating from the Ottoman period that are of Turkic origin or Arabic words that were transmitted via Turkish. A while back, there was a thread about "Albanian-Berber" parallels, and you will notice that many of the words presented there (notably the word for elephant, "fil") are both originally borrowed from Arabic.

The problem is, with such a large share of non-native words (even if they are borrowed from other IE languages), it makes it very hard to exactly pinpoint the position of Albanian inside IE.

I know all this. I have read the Oxford Introduction to Proto-Indo-European as well as most of your posts on the forum. I think that you fail to see what I meant. It doesn't matter how the satemisation of IE pronunciation happened, nor when it happened. The new idea here is that R1a people (represented chiefly by the Balto-Slavic and Indo-Iranian branches, as other branches also have sizeable R1b populations) could not have spoken the same language as R1b people.

Well yeah, but my point is that technically we must assume that the original condition was neither Centum nor Satem, that is, my point is that we should be technically talking about processes of both Centumization (merger of the palatovelars with the plain velars) and Satemization (development of the palatovelars into fricatives).

The fact that the original sound of the palatovelars was a different one to either can be very easily made by the fact that both Centum and Satem languages have plain velars (*k, *g, *gh) and fricatives (*s) corresponding to each other 1:1.

Isolated words in one language or one family (e.g. Western Romance) could indeed be new words coined after the expansion of IE languages. But they could just as well be older loanwords from pre-IE languages. There is no way to know..

It's not quite true that there is "no way to know". If a word makes sense as to be derived from an otherwise known PIE word, then it makes sense to be a newly-coined word.

If we do find similar words that clearly aren't pan-Indo-European and that these words are distributed in several languages in a same region sharing some common non-IE ancestry, then one could wonder whether this word just doesn't come from another (pre-IE) language. This is the point I was trying to make with the Latin vacca, with relatives in other regions with high frequencies of haplogroup J2, both in Indo-European and Semitic languages (that doesn't mean that the word is Semitic if it isn't IE; in this case it is probably from a completely unrelated language that was replaced by Semitic and IE languages).

I see your point. Off hand, there's definitely a few words that come to my mind that exhibit the regional distribution that you had in mind:

*taur- (bull), found in Celtic (Gaulish "tarvos", Irish "tarbh"), Lusitanian ("taurom"), Latin ("taurus"), Greek ("tauros"), Baltic (Lithuanian "tauras", Old Prussian "taurs" - there meaning 'auroch').

*kapr- (goat), found in Celtic (Gaulish "gabro-", Irish "gabhar"), Latin ("capra"), Germanic (Old Norse "hafr" - "he-goat", note that the word exists in many modern Germanic languages but means "oats"), Greek "kapros" (where it actually means "boar", rather than "goat").

What I also should add is that I would like to direct your attention towards the concept/term "wanderwort" (that is, 'wandering word'), which refers to words that are commonly found in various languages languages, but the origin is obscure.

Greek is just as unrepresentative as Latin, if not more. How could Greek language be overwhelmingly Indo-European in origin if only 25% of male lineages in Greece are Indo-European and an even smaller percentage of maternal lineages ? Most Greek R1a being close to the Slavic one, early Middle Ages, and which therefore didn't influence ancient Greek Besides, Indo-European speakers conquered ancient Greece at least twice (Mycenaeans and Dorians), as reflected by the shift from Q-Greek to P-Greek that you explained to me.

It's one thing to hypothesise how languages spread 5000 or 6000 years ago. But it's always good to compare what really happened in historical cases. I do not know any proven historical example of a small minority of people that successfully imposed their language on conquered people and eradicated the native language. There are at least three major obstacles to this in the case of ancient Greece :

1) The Indo-European conquerors that first set foot in Greece, the Mycenaeans, probably didn't make up more than 5% to 10% of the population after the conquest. They were a small ruling elite. So were the Dorians later. They could have imposed their language for administration, but a major native substrate must have survived.

2) the conquered people (pre-IE Greeks) belonged to a quite advanced civilisation for the time. The Indo-Europeans were only superior militarily. Proud civilizations don't give up their language easily to "barbarians".

3) The Indo-Europeans brought a completely new and unrelated language to Greece, which must have been difficult for the natives to learn, even if there was a political will to do so. This argument is well demonstrated by Nicholas Ostler in Empire of the Words. The Romans never succeeded in imposing Latin to the common people in the eastern half of their empire because a) Greek speakers felt too proud/superior to give up their language, and b) Latin was too different and difficult to learn for the uneducated masses of Semitic speakers in the Near East. The Arabs were the only ones who managed to replace local languages in the Levant, Mesopotamia and North Africa because Arabic was sufficiently close to Aramanic, Egyptian, Berber, etc. to be learnt easily by everyone.

My point here is that Greece has a too small percentage of Indo-European ancestry for Greek to be almost purely Indo-European in origins. It's hard to see how more than half of Greek words could be of Indo-European origin. Latin might have a bit more, but it depends if we take Classical Latin or Vulgar Latin.

I would say that with Greek, you have a much better case for pre-IE words than with Latin. There's a large set of words which are clearly pre-Greek. A great example would be the Greek word for "sea", Thalassa, which stands in contrast to the word 'mor-' widely found in Celtic (Gaulish 'mor-', Welsh 'môr'), Germanic (English '-mare', German 'Meer'), Latin ('mare') and Slavic (Russian 'more'). In the past the term "Pelasgian" has been used, but the term has been misused so terribly by people with dubious agendas I prefer to refrain from using it...

I didn't suggest in any way that Italic, Celtic or Germanic languages borrowed words from the ancestral language of Basque. I said from Neolithic languages, of which there were plenty, and many were unrelated or only very distantly related. When we see the number of non-IE West Asian languages that existed until the late Bronze Age (Hurrian, Urartian, Hattian, Kassite, Sumerian, Elamite...), it wouldn't surprising if Neolithic Europe and West Asia had hundreds of languages.

'hundreds of languages' is probably a bit off the mark, but other a large number is absolutely conceivable.

Well, everything you have learnt about Indo-European linguistics was written by people who had zero knowledge of current population genetics. If they build their whole theory on mistaken foundations, there will be a lot to reconsider.

That's a very bold attitude of yours, but I'm curious to see where this project goes... :cool-v:
 
Greek is just as unrepresentative as Latin, if not more. How could Greek language be overwhelmingly Indo-European in origin if only 25% of male lineages in Greece are Indo-European and an even smaller percentage of maternal lineages ?

What is the proof that Mycenean were all R1b or R1a? When they entered Greece, Indo European had already mix with Balkan people. By the way, actual Turkish people have a lot of non Indo European Y dna but Hittite is the closest related language to proto IE.
 
Well, everything you have learnt about Indo-European linguistics was written by people who had zero knowledge of current population genetics. If they build their whole theory on mistaken foundations, there will be a lot to reconsider.

For instance, what does the fact that Basque are 80% R1b tells us about the origin of their language?
In what way does present day Y dna from Turkey explain why Turks (of Turkey) speak a Turkish language?

The new idea here is that R1a people (represented chiefly by the Balto-Slavic and Indo-Iranian branches, as other branches also have sizeable R1b populations) could not have spoken the same language as R1b people.


The Indo europeans mummies of North West China have been tested R1a though Tocharians is a centum language and has similariies with Hittite and the Italo-Celtic language family.

As for R1b, I support that it came separately with the Bell beakers, but I have no idea concerning their languages.
 
What is the proof that Mycenean were all R1b or R1a? When they entered Greece, Indo European had already mix with Balkan people. By the way, actual Turkish people have a lot of non Indo European Y dna but Hittite is the closest related language to proto IE.

There are two good reasons to think that the Mycenaeans were mostly (not exclusively) R1a and/or R1b :

1) The Indo-Europeans from the Pontic Steppes first made incursions to the Balkans, and all of them passed through the Balkans (this wasn't exactly a fast migration, it took several centuries) before reaching Central and Western Europe. Yet, even after centuries since the first IE cultures (Cotsofeni, Cernavoda, Glina, Otomani, etc.) in the Balkans the vast majority of Indo-Europeans who reached Western Europe were R1b. I don't see why the migration to Greece, which was closer to the Pontic Steppes, would be any different. If the Mycenaeans did come through this early expansion of the southern Yamna culture, then all the chances are that they belonged overwhelmingly to R1b.

2) Artefacts from the Mycenaean Culture bore striking resemblance to the Kurgan steppe culture, which suggest that they might have moved directly from the steppes to Greece. If that is so, they would have been principally R1a. This is the hypothesis I originally supported due to the late commencement of the Mycenaean Culture (from circa 1650 BCE, contemporary to the Timber-grave Culture in the steppes) and the close kinship of Mycenaean language to Proto-Indo-Iranian.

So I am still in two minds as to the dominant haplogroup of the Mycenaeans. We definitely need more detailed subclade testing for Greek R1a and R1b. If the Mycenaean were R1a, then they must have made up less than 5% of the population since a big part of the 15% of modern Greek R1a is Slavic or Turkic.

R1b is also problematic since most of the Greek R1b is the so-called ht35 (R1b-L23), which is most probably pre-IE from Anatolia. Greeks also have a lot of R1b-S28, which could be Roman or Celtic (the Dorians could have been Central European Celts who brought the Q/P shift). If the Mycenaeans were Yamna R1b people, we should expect to find a considerable portion of R1b-L51 (the most common in Romania and Bulgaria, representing the first wave of R1b out of the steppes) and R1b-L11.
 
So I am still in two minds as to the dominant haplogroup of the Mycenaeans. We definitely need more detailed subclade testing for Greek R1a and R1b. If the Mycenaean were R1a, then they must have made up less than 5% of the population since a big part of the 15% of modern Greek R1a is Slavic or Turkic.

I would be a bit cautious with "Turkic" R1a. How much of the present-day R1a in Turkey is genuinely Turkic, and how much stems from pre-Turkic (Greeks, Persians, etc.) or otherwise non-Turkic (Slavic slaves during the Ottoman period, for instance), and how much, if at all, R1a did the Turks actually bring with them from Central Asia? I think that the situation in Anatolia when it comes to this is a lot more complex.

R1b is also problematic since most of the Greek R1b is the so-called ht35 (R1b-L23), which is most probably pre-IE from Anatolia. Greeks also have a lot of R1b-S28, which could be Roman or Celtic (the Dorians could have been Central European Celts who brought the Q/P shift). If the Mycenaeans were Yamna R1b people, we should expect to find a considerable portion of R1b-L51 (the most common in Romania and Bulgaria, representing the first wave of R1b out of the steppes) and R1b-L11.

The interesting part with the Dorians is that while there are good arguments from both genetics and linguistics that the Dorian invasion really happened, the archaeological situation is highly contested at best.
 
R1b is also problematic since most of the Greek R1b is the so-called ht35 (R1b-L23), which is most probably pre-IE from Anatolia. Greeks also have a lot of R1b-S28, which could be Roman or Celtic (the Dorians could have been Central European Celts who brought the Q/P shift). If the Mycenaeans were Yamna R1b people, we should expect to find a considerable portion of R1b-L51 (the most common in Romania and Bulgaria, representing the first wave of R1b out of the steppes) and R1b-L11.
I think this is a telling point: The L11 in Greece is extremely sparse, so I would say the Mycenaeans were most probably R1a.
 
For instance, what does the fact that Basque are 80% R1b tells us about the origin of their language?

The Basques are an exception among R1b populations.

In what way does present day Y dna from Turkey explain why Turks (of Turkey) speak a Turkish language?

It's very hard to tell exactly how much Turkic DNA there is in Turkey because Turkic speakers from Central Asia were extremely heterogeneous. The obvious haplogroups are the original Northeast Asian ones (C, N, O, Q, accounting for about 8% of the haplogroups in Turkey), but Central Asians also have a very considerable amount of R1a, and reasonably lots of G, J2 and R1b as well, three haplogroups which are very common in Turkey. There could be up to 20-25% of haplogroups of Turkic/Central Asian origin in Turkey, or as little as 10-15%.

The Indo europeans mummies of North West China have been tested R1a though Tocharians is a centum language and has similariies with Hittite and the Italo-Celtic language family.

What makes you think that the tested Tarim mummies (2000-1800 BCE) are related to Tocharian language (circa 300-900 CE) ? There is over 2000 years separating them. I personally think that the Tarim mummies were Indo-Iranian speakers, because they appeared just at the time of the Indo-Iranian expansion in Central Asia. Tocharian speakers probably came much later, and I would expect them to be R1b-M73 since 20% of modern Uyghurs belong to that haplogroup and it wasn't found among Tarim mummies. The last Tarim mummies (3rd century CE) coincide with the arrival of Tocharian speakers, just like the disappearance of Tocharian language coincides with the arrival of Turkic-speakers (9th century).

As for R1b, I support that it came separately with the Bell beakers, but I have no idea concerning their languages.

And where did the Bell bekaers come from ?
 
The Basques are an exception among R1b populations.

Today, yes. We however don't know if the Iberians were predominantly R1b, too, for instance, but it's likely.

And where did the Bell bekaers come from ?

Well, we had the discussion about Beaker-Bell and it's affiliation with Indo-European and/or R1b before. The oldest metal-working sites of the Beakers undoubtably are in western Portugal, so one could speculate if Beaker-Bell started out as a native phenomenon (unless you believe the stelae people hypothesis, which in my opinion has some serious problems).

I must confess there is no satisfying answer to Beaker-Bell yet.

What makes you think that the tested Tarim mummies (2000-1800 BCE) are related to Tocharian language (circa 300-900 CE) ? There is over 2000 years separating them. I personally think that the Tarim mummies were Indo-Iranian speakers, because they appeared just at the time of the Indo-Iranian expansion in Central Asia. Tocharian speakers probably came much later, and I would expect them to be R1b-M73 since 20% of modern Uyghurs belong to that haplogroup and it wasn't found among Tarim mummies. The last Tarim mummies (3rd century CE) coincide with the arrival of Tocharian speakers, just like the disappearance of Tocharian language coincides with the arrival of Turkic-speakers (9th century).

Sorry Maciamo, I'm afraid that really makes no sense. Where would the Tocharians come from if they only arrived in the 3rd century AD? No matter which hypothesis you prefer in regard for the affiliation of the Tocharian language, it's clear that the language must be pretty ancient, and 2000 BC is indeed a likely date for the arrival. In my opinion, the Tocharians were indeed R1a, and M73 is a Haplogroup of the Uighurs/Turkic peoples.

I also don't think that there were Indo-Iranic peoples in the Tarim basin. Mind you, before the discovery of the Tocharian inscriptions, nobody even suspected Indo-Europeans to have lived there.
 
And where did the Bell bekaers come from ?

From the east as a lot of them were Bracycephal. When we look at the actual distribution of R1b in France, the area (Inner France) not settled by the Bell Beaker has more Neolitic haplogroup like E1b1, J1
 
Genetics is a useful addition to historical, archaeological, and linguistic explanations of ethnic and political processes. That's all. They are certainly not a key that opens all doors or perhaps even many doors. By the time history begins genetic relationships were already pretty complicated. I had to laugh at the fairy tale that R1b and R1a people shared the same territory and culture but did not speak the same language or even interrelate genetically.It's certainly great to own your own site. You can create your own fantasies with impunity. Oh well one has to take the bad with the good I guess. And there's lots of good things here as well as incredible hogwash.
 
Genetics is a useful addition to historical, archaeological, and linguistic explanations of ethnic and political processes. That's all. They are certainly not a key that opens all doors or perhaps even many doors. By the time history begins genetic relationships were already pretty complicated. I had to laugh at the fairy tale that R1b and R1a people shared the same territory and culture but did not speak the same language or even interrelate genetically.It's certainly great to own your own site. You can create your own fantasies with impunity. Oh well one has to take the bad with the good I guess. And there's lots of good things here as well as incredible hogwash.

I never said that R1a and R1b people shared the same territory at the same time. I always insisted that during the Yamna horizon, R1a people lived in the northern forest-steppe secluded from the R1b people in the southern grassy steppe and the North Caucasus. I think that a minority of R1b people invaded the R1a people to the north and brought with them the bronze technology (from the Maykop Culture, the oldest known bronze age culture in the world), which prompted the Corded Ware expansion to the north, northwest and east. Unfortunately, once the bronze technology acquired, the R1a people tried to expand southward as well, which caused tensions with the R1b populations there. R1b tribes started pushing their way to the Balkans attracted by the copper and gold, and pressed by their northern R1a neighbours. R1b tribes eventually left the Pontic steppes completely and migrated along the Danube to Central and Western Europe. By the time of the Catacomb Culture (2500 BCE), the (grassy) Pontic Steppes had become entirely settled by the northern R1a, who filled the vacuum. The R1a expansion continued eastward to Central and South Asia from 2000 BCE. The two groups were always distinct, except for the initial foray of R1b to the Forest-Steppe (between 4000 and 2900 BC).

I am sorry for you that this is beyond your intellectual capacities to comprehend.
 
*kapr- (goat), found in Celtic (Gaulish "gabro-", Irish "gabhar"), Latin ("capra"), Germanic (Old Norse "hafr" - "he-goat", note that the word exists in many modern Germanic languages but means "oats"), Greek "kapros" (where it actually means "boar", rather than "goat").

The Centum root for goat is obviously *kapr- so I wonder why Mallory & Adams list 8 other PIE roots for goat. Goat is koza in all/most Slavic languages, kaza in Latvian and ozka in Lithuanian. Only one root here too, though a different one from the Centum languages. The Sanskrit word is aja, which was replaced by kamuk or bakri in Hindi. All three are probably of non-IE origin (unless bakri is a dyslexic inversion of kapri, like the Lithuanian ozka from koza).

I don't know where the other roots I listed above come from (Albanian, Dacian, Thracian, Armenian, Tocharian, Hittite ?) but it's unlikely that there are more than two PIE roots. Goat was the Neolithic animal par excellence (along with cattle), so in all likelihood local communities all over Europe, the Middle East and South Asia kept older words for it after the Indo-Europeans arrived. I noticed that words for domesticated animals tend to stick more than others after a change of language. Walloon is a Romance language, yet the word for goat is gat (Germanic) and the word for sheep is bedo, which is non-Indo-European.
 
I have created a new thread regarding the question of the connection between the Tocharians and the Tarim mummies as we were getting off-topic and such an interesting subject deserves its own thread.
 
The Centum root for goat is obviously *kapr- so I wonder why Mallory & Adams list 8 other PIE roots for goat. Goat is koza in all/most Slavic languages, kaza in Latvian and ozka in Lithuanian. Only one root here too, though a different one from the Centum languages. The Sanskrit word is aja, which was replaced by kamuk or bakri in Hindi. All three are probably of non-IE origin (unless bakri is a dyslexic inversion of kapri, like the Lithuanian ozka from koza).

I don't know where the other roots I listed above come from (Albanian, Dacian, Thracian, Armenian, Tocharian, Hittite ?) but it's unlikely that there are more than two PIE roots. Goat was the Neolithic animal par excellence (along with cattle), so in all likelihood local communities all over Europe, the Middle East and South Asia kept older words for it after the Indo-Europeans arrived. I noticed that words for domesticated animals tend to stick more than others after a change of language. Walloon is a Romance language, yet the word for goat is gat (Germanic) and the word for sheep is bedo, which is non-Indo-European.

If it is of any help to compare, the albanian word for 'goat' is 'dhi', 'male goat' is 'cjap' and for 'sheep' is 'dele' and 'male sheep' is 'dash'.
 
If it is of any help to compare, the albanian word for 'goat' is 'dhi', 'male goat' is 'cjap' and for 'sheep' is 'dele' and 'male sheep' is 'dash'.

Thanks. These are probably pre-IE words.
 
The Centum root for goat is obviously *kapr- so I wonder why Mallory & Adams list 8 other PIE roots for goat. Goat is koza in all/most Slavic languages, kaza in Latvian and ozka in Lithuanian. Only one root here too, though a different one from the Centum languages. The Sanskrit word is aja, which was replaced by kamuk or bakri in Hindi. All three are probably of non-IE origin (unless bakri is a dyslexic inversion of kapri, like the Lithuanian ozka from koza).

I don't know where the other roots I listed above come from (Albanian, Dacian, Thracian, Armenian, Tocharian, Hittite ?) but it's unlikely that there are more than two PIE roots. Goat was the Neolithic animal par excellence (along with cattle), so in all likelihood local communities all over Europe, the Middle East and South Asia kept older words for it after the Indo-Europeans arrived. I noticed that words for domesticated animals tend to stick more than others after a change of language. Walloon is a Romance language, yet the word for goat is gat (Germanic) and the word for sheep is bedo, which is non-Indo-European.

If it is of any help to compare, the albanian word for 'goat' is 'dhi', 'male goat' is 'cjap' and for 'sheep' is 'dele' and 'male sheep' is 'dash'.

I'd like to shed light on two words in the list here:

Albanian "cjap" (note that Albanian "c" is pronounced as "ts") has a cognate in Romanian "țap-", and hence can be traced to a common root *tsap-.

Furthermore, Sanskrit "aja" has cognates in Greek and Armenian. Bear in mind that "kapros" means 'boar', and an actual word for 'goat' is "aiga". In a similar fashion, Armenian has the word "aits". If we consider Indic and Armenian sound laws, the ancestral form can be faithfully reconstructed as *aig´-.
 

This thread has been viewed 42233 times.

Back
Top