European languages before Indo-European

[Tropylium]

I'll add a few then:

Thank you! Quite interesting.

I don't think this is sufficient evidence to make the change "common as dirt", though.

It seems to me that precedents for a merger of velars and uvulars are actually difficult to find.

But this assumes that the distinction was velar-uvular at the time of the merger, rather than say palatal-velar or palatalisedvelar-velar or frontvelar-backvelar.

Not really. The problem of finding precedents gets even worse if you start with a more traditional model of PIE, which is what has of course motivated the newer velar-labiovelar-uvular model in the first place.

And then of course from the other point of view, a theory of centum languages as the result either of universal substratum influence or as the result of a shared development in a subnode must deal with the problems of Anatolian and Tocharian.

That is not the only other option, though. Other possibilities include e.g.
1) there were only plain and labiovelars in PIE; satemization is initiated very early; all words with "plain velars" are either the result of neutralization (*ḱr > *kr, etc.) or post-PIE loanwords into the Satemic varieties
2) the front/back velar contrast existed in PIE but was carried on something other than the consonants, and satemization/centumization were conditional rather than unconditional sound changes (e.g. *ke > *ḱe, but *kæ > *ke).
(But this is getting into PIE Thread territory, of course.)

[KathTheDragon]

I'll add a few then:

Thank you! Quite interesting.

I don't think this is sufficient evidence to make the change "common as dirt", though.

Not "common as dirt", but certainly not unusual or even improbable. But then again, since when does a change have to be "common as dirt" in order to be postulated for the development of a protolanguage?

[WeepingElf]

I'll add a few then:

Thank you! Quite interesting.

I don't think this is sufficient evidence to make the change "common as dirt", though.

OK, it is not "common as dirt" then. But I still think it is not out of the ordinary.

It seems to me that precedents for a merger of velars and uvulars are actually difficult to find.

But this assumes that the distinction was velar-uvular at the time of the merger, rather than say palatal-velar or palatalisedvelar-velar or frontvelar-backvelar.

Not really. The problem of finding precedents gets even worse if you start with a more traditional model of PIE, which is what has of course motivated the newer velar-labiovelar-uvular model in the first place.

Fair. I am not sure which model is the best, hence I prefer operating with the neutral terms "front velar" and "back velar" here. Also, IMHO the labiovelars were labialized back velars.

My hypothesis that the three velar series preserve vowel features [+front] and [+round] that were lost in the Great Vowel Collapse supposes that the front velars were palatalized velars and the back velars plain velars. This system would have been notoriously unstable as the least marked series was the least common (though there are precedents for that, in the Caucasus and on the North American Pacific coast, for instance), hence it "normalized" in two directions - by losing palatalization in the centum dialects and by palatalizing the front velars and losing labialization in the satem dialects.

And then of course from the other point of view, a theory of centum languages as the result either of universal substratum influence or as the result of a shared development in a subnode must deal with the problems of Anatolian and Tocharian.

That is not the only other option, though. Other possibilities include e.g.
1) there were only plain and labiovelars in PIE; satemization is initiated very early; all words with "plain velars" are either the result of neutralization (*ḱr > *kr, etc.) or post-PIE loanwords into the Satemic varieties
2) the front/back velar contrast existed in PIE but was carried on something other than the consonants, and satemization/centumization were conditional rather than unconditional sound changes (e.g. *ke > *ḱe, but *kæ > *ke).

I can't say that it can't be the case, and there are indeed Indo-Europeanists who favour the first hypothesis (don't know about the second, though).

(But this is getting into PIE Thread territory, of course.)

Indeed it does.

[Tropylium]

OK, it is not "common as dirt" then. But I still think it is not out of the ordinary.

Probably not that either, no. My bad if "[citation needed]" came off as too belligerent.

(But this is getting into PIE Thread territory, of course.)

Indeed it does.

In an attempt to steer back on course, I could note that uvular/velar contrasts seem to be also strongly areal. The Old World has one large zone of them, covering Central Asia and Siberia, and reaching thru Yukaghir and Chukotko-Kamchatkan (but not Tungusic, for the most part) to NW America; a medium-size one covering the Caucasus and connecting thru Semitic to northern Africa; and a third smaller group in southern China, possibly historically connected to Austronesian. In the New World, they stretch quite a bit down south along the Pacific coast, and reappear further south yet along the Andes, but remain mostly absent in the cismontane US (is this a term?) and the Amazon.

This suggests not only that a common substrate could be blamed for centumization; but also that if we assumed a merger as a native innovation, it should be quite possible for this to overtake large swaths of western IE varieties.

[Nortaneous]

I don't have the time to look now, but I'd look at Hmong-Mien for more examples of sound changes affecting uvulars -- might be some mergers with velars there. PHM had uvulars and Iu Mien doesn't, so who knows.

[Salmoneus]

I think it's worth pointing out the methodological problem of judging potentially 'common as dirt' changes. If a change IS so common, or so easy, that it can occur many times independently, then it is only to be expected that there is relatively little evidence of it.

Discounting languages with long written records (and even there some contrasts, like velar-uvular, may not be reflected in writing, particularly if the script is non-native), we can only guess at a former velar-uvular contrast where two related languages have different outcomes - ideally where the contrast remains in at least one language, and we can maybe guess at it from other processes too (like satemisation vs centumisation - but note that even here we'd have no reason to even suspect uvulars if PIE didn't also clearly have labiovelars - otherwise we'd probably just write it off as a velar-labiovelar contrast).

But if the contrast is something that is very likely to neutralise, then there's a high change that there won't be enough remnants of it left for us to know it was there in the first place. We're left only knowing about those uvulars that didn't just merge with velars quickly.

[jal]

Interesting article on the roots of Europeans here.


JAL

[Tropylium]

I think it's worth pointing out the methodological problem of judging potentially 'common as dirt' changes. If a change IS so common, or so easy, that it can occur many times independently, then it is only to be expected that there is relatively little evidence of it.

Good point. True in principle, but this does not mean we have an open cheque to claim that any random change such as ⁿǀ > m or ħʷ > j is common a dirt, and therefore we should be allowed to reconstruct *ⁿǀ or *ħʷ for dozens of language families despite the complete absense of explicit evidence. There has to be also some initial evidence that is available from historical records, narrow dialectal or idiolectal variation, synchronic (morpho)phonological processes, and the like.

(The first of these is the most reliable, but also the most limited; the second and third provide more examples, but will not tell us much about directionality, and they can sometimes indicate structures with a common origin rather than structures that have actually changed into one another.)

[Salmoneus]

Just came here to talk about [EDIT: jal's link], having read through it and looked at some commentaries.

The headline news is that the theory I mentioned before about Yamnaya having Caucasian ancestry is correct.

A quick run-down of what I now understand to be the case:
1. Around 45,000 years ago, people entered Europe. This population split off from the other Eurasian groups at this point. East Asians split off at a somewhat similar time, I think. [Of course, this is just when THESE people entered Europe. It's possible that other Europeans had already gone there, but later got wiped out]

2. A bit before 20,000 years ago, people in the ME region split in two. This may be related to the ice age: one group lived south of the mountains, in the Levant, while another group lived north of them, in modern Georgia, or perhaps the southern coast of the Caspian.

3. Caucasians and Europeans were very inbred. Western Europeans across thousands of years and hundreds of miles were closely related, and they also showed an extremely high degree of consanguinity (ie incest) in their genes. In this respect they're similar to modern Austronesian or Andaman populations: there may not have been literal incest, but we're talking small, isolated communities. At the same time, those communities were related across time and space. As in, substantially more closely related than modern Europeans are to one another. The paper Jal links to mentions that the common origin of a 14kya European from Neuchatel in Switzerland and an 8kya European from Luxembourg was only 16kya.
Similarly, the Caucasian samples suggest inbreeding - a relatively small population. The Levantines, on the other hand, seem to have had a larger population to draw their genes from. Again, the ice age may be responsible. Although Europe and the Caspian shore weren't frozen, I guess they would have been pretty cold, barren and uninviting. Populations living there probably wouldn't have been big, and in the case of Europe may have had to move around a bit.
Looking deeper into the notes: the Caucasian and European samples all suggest population bottlenecks: expansion from a small number of ancestors. The older Caucasian sample also suggests recent inbreeding within the family - that may be coincidence, or culture, or may just indicate a particularly small community at that time.

4. There was another population in India, related to modern Andaman Islanders. I gather than the andaman Ongan people, the Papuans/pre-Melanesians and the Australians are all racially connected to one another, and to various 'Negrito' populations that have since been overrun by east asian linneages (Sino-Tibetan, Austro-Tai).

5. People in the Levant invented farming and invaded Europe. These people almost obliterated 'Old European' linneage in southern Europe, but more of the old lines survived further north and west.

6. At a similar time, Caucasians expanded too. Presumably because they'd picked up farming from their southerly neighbours. They probably spread across Iran and Central Asia, and down into India, where they interbred with the local Ongans. The result of this interbreeding was probably the Dravidian peoples. [Although of course the Dravidian languages might either pre- or post-date this Caucasian invasion].

7. They also spread north onto the Steppe. The Maikop culture appears to have spread Caucasian genes north of the Caucasus, and to have played a founding role in the creation of the Yamnaya culture. The Yamnaya were approximately 50% Caucasian, and 50% 'Eastern Hunter Gatherer'. These EHGs were most closely related to the Old European hunter-gatherers, but they also had an element shared with Siberia. So basically, there were probably subglacial hunter-gatherers independently emerging in Europe and Siberia, and in a region from eastern europe to the western steppe these two groups merged. Now interestingly, I seem to gather that in the Yamnaya the Caucasian element may have been mostly maternal. This may suggest pastoralists invading their neighbouring farmers and stealing their wives, in which case the Yamnaya language (PIE) probably wouldn't be Caucasian... or, alternatively, it may just show some pattern of acquisition of prized Maikop women by ruling Yamna men, whether through trade or diplomacy. Or, of course, the pre-Yamna may have been a matriarchal culture and the women just really got turned on by big dark hunter-gatherer men. Probably not, though.

8. The Yamnaya then invaded both Europe (where they replaced the Middle Europeans of the LBK etc) and India (where they mingled with the comingled Ongan-Caucasians, and formed a ruling caste. In Europe, the replacement was more severe in the north than in the south, so Caucasian genes are found more commonly in the north.

9. Meanwhile, the old Levantines were invaded by Africans. This is presumably the influx of Afroasiatics. So it's the same pattern as with the Yamnaya: first a spread of farmers, then a spread of pastoralist conquerors. I wonder if the proto-Sino-Tibetans were also pastoralists who displaced the agriculturalist Austro-Tai/Hmong-Mien?

10. The Uralics don't show up. They look pretty much the same as the Yamnaya. That might indicate they were the same people (Indo-Uralic). Or it may mean they were neighbours. Or just that wherever Uralic languages have spread, they've spread to Yamna populations and haven't dented the genes.

There are also a couple of interesting things when it comes to modern populations:
- Scotland stands out with a rogue 'west asian' gene contribution. Don't know what that means.
- the North Caucasus, and a few Indo-Iranian spots, show up with "American" genes.
- Cyprus shows up with Caucasian genes, although that might be some freak of later IE migration.

Oh, and
11. The Levantine farmers probably had two parent populations. One were "Basal Eurasians", the same people who produced Caucasians and East Asians. The others were 'West Eurasians', who also produced the oldest Europeans. West Eurasians in turn are siblings of North Eurasians, who produce the siberian hunter-gatherers. Together, those two groups are the sibling to the group that ultimately produced the Ongans/Papuans, and THAT pairing is sibling to the Basal Eurasians. But I think this very-long range sort of discussion is more tendentious without more old samples.

---------------

What does this mean for Mesolithic Europe?
Well, I think the likely picture here is not a rich tapestry of independent languages, or not exactly. Yes, they could have had 40,000 years to develop a range of languages. But the evidence of population bottlenecks, small populations, an unwelcoming environment and relatiely close genes even as late as the mesolithic, combined with the possibilities of post-glacial expansions, all suggest much more similarity than that.

So here's my scenario:
- A Gravettian culture is widespread throughout the plains. The Gravettians are the genetic West/North Eurasians - the ancestors of the European, Siberian, and American hunter-gatherers.
- Climates deteriorate, and that unity is broken up. Western and Northern/American hunters are separated. Some Gravettians in the Balkans or Anatolia interbreed with a Southwest Asian stock to produce the people who will eventually develop farming.
- In Europe, a population is bottlenecked in southwest France: the Solutreans
- As the climate begins to change around 17,000, the Solutreans expand.

Now, interestingly, the old europeans genetically furthest from the jura/luxemburger samples are the Scandinavians, while the Hungarians and the Spanish are about equally distant from them. This suggests we're not talking gradual spread across the continent, but rather a quick split into groups. On the other hand, these were probably hunters with seasonal ranges, so they moved around a lot and probably interbred with one another, explaining the genetic closeness. So:

- groups of Solutreans, now called the Magdalenians, push north into Scandinavia and become cut off, while others turn south into Pannonia and become cut off. Others may head further east, interbreeding with siberians they meet, although apparently less was happening in the east. And others turn back into Spain and become cut off by the Pyrenees. The Magdalenian culture never makes it to Italy, however, nor the Balkans - these are conservative 'epigravettian' cultures instead, probably because these areas are repopulated by non-solutreans.

The Magdalenians, it should be noted, probably were assisted by new bone-carving technologies and the domestication of the dog, which might explain why they exploded faster than their rivals to the east.

--------------
-----------------

So how many language stocks were there?

Early Palaeolithic: everyone would have been related, but not necessarily all that closely. We're probably looking at an immense sprachbund of semi-nomadic hunters, with a common language but tens of thousands of years of variation. Then again, semi-nomadic, perhaps reminiscent of the plains of NA... and the material cultures remained similar. We're probably talking all related languages, but with connections obscured by areal effects and diversification.

Later Palaeolithic: ultimately but obscurely related stocks in the Southeast, in Italy, and Everywhere Else (Magdalenian). The Magdalenian languages were probably divided into northern, eastern, southwestern and central language groups - within the group there may have been distinct languages, or just dialect continua, like in Australia. Meanwhile the Italian and Balkan areas probably have one big language group each, since the founder populations would have been small.

Mesolithic: in the Mesolithic, the cultural continuity breaks down, and people become less mobile, so probably language continuity breaks down. At this point you may have half a dozen language families, each divided into substantial branches. Eg. the Central Magdalenian family with southwest french, northern french, british, low countries, german etc families. Perhaps equivalent to IE branches?

Neolithic: everything except the atlantic coast (and maybe scandinavia, scotland, the baltic, etc) wiped out by near eastern languages. there would have been two great language families (LBK and Cardial) which may themselves have been related (would have been related, really - but maybe not for ten thousand years, or maybe one thousand years, no way to tell). Basque may have been a pre-neolithic (ie Magdalenian) family, or may have been just a slow-to-arrive neolithic language.

--------

*tired now*

[WeepingElf]

Interesting post, Salmoneus. Nice to see an alternative opinion to my admittedly rather naive guess of "20-30 stocks". At least, you have some reasonable argumentation behind your estimate, even if genes and languages spread independently of each other, with languages spreading much faster. And if two languages are related by a common ancestor spoken 40,000 years ago, this relationship is probably pretty much impossible to discern now, as everything has changed beyond recognition in those languages. "Vasco-Caucasian" may be real - but undemonstrable. You address the matter of Sprachbünde, which certainly is something to consider.

[hwhatting]

"American" genes in the North Caucasus - that looks like a boost to the Dene-Caucasian hypothesis...
Totally different question - WeepingElf, was it you who talked here some time ago about a substrate language with ai-a and au-a ablaut? If yes, do you have some sources? Was that one of Kuhn's ideas? I'd like to look a bit into the material for that.

[WeepingElf]

"American" genes in the North Caucasus - that looks like a boost to the Dene-Caucasian hypothesis...
Totally different question - WeepingElf, was it you who talked here some time ago about a substrate language with ai-a and au-a ablaut? If yes, do you have some sources? Was that one of Kuhn's ideas? I'd like to look a bit into the material for that.

I wasn't the one with the "ai-a and au-a ablaut". No such thing figures in any of my hypotheses. I don't know who else posted that, either.

As for the degree of diversity, Sal and I may be talking past each other as it is far from obvious how to count language families. Surely, the notion of deepest established units does not apply here as we are dealing with unknown languages. When I talked about "20-30 stocks", I meant stock to mean a language family with a time depth in the range of 4,000 to 6,000 years. For instance, in the present day, Indo-European, Uralic and Semitic are stocks. This does not exclude the possibility of those stocks being related to each other. Indeed, Afroasiatic is a group of six stocks that are widely accepted to be related to each other; and IE and Uralic are IMHO probably related to each other as well.

[hwhatting]

"American" genes in the North Caucasus - that looks like a boost to the Dene-Caucasian hypothesis...
Totally different question - WeepingElf, was it you who talked here some time ago about a substrate language with ai-a and au-a ablaut? If yes, do you have some sources? Was that one of Kuhn's ideas? I'd like to look a bit into the material for that.

I wasn't the one with the "ai-a and au-a ablaut". No such thing figures in any of my hypotheses. I don't know who else posted that, either.

Thanks anyway. Anyone else knows? Otherwise I'll just search the depths of the internets...

[Salmoneus]

Interesting post, Salmoneus. Nice to see an alternative opinion to my admittedly rather naive guess of "20-30 stocks". At least, you have some reasonable argumentation behind your estimate, even if genes and languages spread independently of each other, with languages spreading much faster. And if two languages are related by a common ancestor spoken 40,000 years ago, this relationship is probably pretty much impossible to discern now, as everything has changed beyond recognition in those languages. "Vasco-Caucasian" may be real - but undemonstrable. You address the matter of Sprachbünde, which certainly is something to consider.

I wouldn't say that genes and languages spread independently. In most cases, they seem to spread together. I would instead say that languages and genes CAN spread independently in some cases. Although I'd suggest that these anomalies are probably less likely the further we go back: language spread requires cultural influence, and it's probably harder and harder to have cultural influence at a distance without major population movements when you're dealing with less and less dense populations with less and less sophisticated forms of transport and communication.

I also wouldn't say that languages spread 'much faster' than genes. What are your examples for that? In Europe, for instance, we have examples of both sorts of anomaly. On the one hand, the Sardinians are nearly pure Neolithic genetically, but they speak a language family that didn't arrive on their island until the iron age. So language has spread faster than genes - sort of. Although there ARE indoeuropean genes on Sardinia, so this ISN'T an example of language spread without genetic spread, just of language overturning without genetic overturning - it's a case of a conqueror language imposed by a minority onto a majority.
But then you've got the Basques, who are genetically a mixture of Neolithic and Bronze Age genes, the same as everybody else, but with a slight but significance influence from Mesolithic genes. Yet they speak a language that is at the very least from a family that's been there since the neolithic, and more likely probably from the mesolithic. So here we have a case where genes have spread faster than language.

----

Regarding stocks: my point was that Mesolithic stocks would derive from probably only three Palaeolithic stocks, or maybe four. So by 5000BC, the languages of France and Germany and Sweden, etc, would be separated by at most 8-10,000 years. That would be more than IE, but maybe similar to Niger-Congo's time-depth, and probably considerably less than the time depth of Afro-Asiatic. That's the worst-case scenario: it assumes that there weren't other large-scale population replacements between the Magdalenian and the late Mesolithic, which wouldn't be a great surprise given the low population size and the development of several new cultures and technologies in the interim.

[Zaarin]

"American" genes in the North Caucasus - that looks like a boost to the Dene-Caucasian hypothesis...

Is that really surprising? Genetic evidence already shows American ancestry in Central Asia/Transcaucasia, and linguistically Dene-Yeniseian seems very promising.

[hwhatting]

"American" genes in the North Caucasus - that looks like a boost to the Dene-Caucasian hypothesis...

Is that really surprising? Genetic evidence already shows American ancestry in Central Asia/Transcaucasia, and linguistically Dene-Yeniseian seems very promising.

To me, yes, as I've never much looked into the genetic side of things. And on Dene-Caucasian I'm just an interested onlooker, as I've almost no knowldege of the languages grouped into that macrofamiliy.

[WeepingElf]

I wouldn't say that genes and languages spread independently. In most cases, they seem to spread together. I would instead say that languages and genes CAN spread independently in some cases. Although I'd suggest that these anomalies are probably less likely the further we go back: language spread requires cultural influence, and it's probably harder and harder to have cultural influence at a distance without major population movements when you're dealing with less and less dense populations with less and less sophisticated forms of transport and communication.

Yes, you are correct. Certainly, most people get their L1 from (one of) their germane parents. There are exceptions, and these must not be neglected, though most of them presuppose a stratified society where the élite and the lower classes speak different languages, and probably did not play a major role before the Neolithic.

I also wouldn't say that languages spread 'much faster' than genes. What are your examples for that? In Europe, for instance, we have examples of both sorts of anomaly. On the one hand, the Sardinians are nearly pure Neolithic genetically, but they speak a language family that didn't arrive on their island until the iron age. So language has spread faster than genes - sort of. Although there ARE indoeuropean genes on Sardinia, so this ISN'T an example of language spread without genetic spread, just of language overturning without genetic overturning - it's a case of a conqueror language imposed by a minority onto a majority.
But then you've got the Basques, who are genetically a mixture of Neolithic and Bronze Age genes, the same as everybody else, but with a slight but significance influence from Mesolithic genes. Yet they speak a language that is at the very least from a family that's been there since the neolithic, and more likely probably from the mesolithic. So here we have a case where genes have spread faster than language.

Fine. Language spread without genetic spread is very rare if it occurs at all, but there are situations where a genetic minority impose their language on a genetic majority. In Sardinia, this linguistically triumphant minority were the immigrants from mainland Italy; in the Basque country, it were the locals.

Regarding stocks: my point was that Mesolithic stocks would derive from probably only three Palaeolithic stocks, or maybe four. So by 5000BC, the languages of France and Germany and Sweden, etc, would be separated by at most 8-10,000 years. That would be more than IE, but maybe similar to Niger-Congo's time-depth, and probably considerably less than the time depth of Afro-Asiatic. That's the worst-case scenario: it assumes that there weren't other large-scale population replacements between the Magdalenian and the late Mesolithic, which wouldn't be a great surprise given the low population size and the development of several new cultures and technologies in the interim.

I see. As I said a few days ago, I no longer maintain the "20-30 stocks" opinion, which was little else than naively applying North American diversity patterns to Europe, and assuming that each of the three major peninsulas of southern Europe had a diversity comparable to the Caucasus. I now realize that it is not quite as simple. Europe is not "half-size North America", it is a region with its very own characteristics, and a very different human prehistory. The Caucasus is a region where probably language families from a larger area were pushed up the mountain valleys by a rising tide of Indo-European and Turkic languages.

I now assume one stock on each of the three great peninsulas, and perhaps two or three stocks north of the Alps in the Mesolithic. Perhaps a little more.

[Zaarin]

"American" genes in the North Caucasus - that looks like a boost to the Dene-Caucasian hypothesis...

Is that really surprising? Genetic evidence already shows American ancestry in Central Asia/Transcaucasia, and linguistically Dene-Yeniseian seems very promising.

To me, yes, as I've never much looked into the genetic side of things. And on Dene-Caucasian I'm just an interested onlooker, as I've almost no knowldege of the languages grouped into that macrofamiliy.

While not an expert on either group, I do have an interest in the Na-Dene languages (particularly Tlingit). Personally, I find Dene-Caucasian a little dubious but Dene-Yeniseian probable--but if a link could be demonstrated between Yeniseian and a Caucasian language...

[hwhatting]

While not an expert on either group, I do have an interest in the Na-Dene languages (particularly Tlingit). Personally, I find Dene-Caucasian a little dubious but Dene-Yeniseian probable--but if a link could be demonstrated between Yeniseian and a Caucasian language...

Well, from what I've read about Dene-Yeniseian, it looks like it has been established in a methodologically sound way. For anything further, I reserve judgment.

[Zaarin]

While not an expert on either group, I do have an interest in the Na-Dene languages (particularly Tlingit). Personally, I find Dene-Caucasian a little dubious but Dene-Yeniseian probable--but if a link could be demonstrated between Yeniseian and a Caucasian language...

Well, from what I've read about Dene-Yeniseian, it looks like it has been established in a methodologically sound way. For anything further, I reserve judgment.

That seems to be the general mood towards Dene-Yeniseian among experts on both sides: cautiously optimistic.

[Tropylium]

Certainly, most people get their L1 from (one of) their germane parents.

This is already sufficient to allow for language shift, FWIW. It's relatively common for parents to raise their children in a language they are not native speakers of; perhaps not even fluent in. Gaps will be filled in as long as the kids have some L1 speakers around to interact with. (And then there's multilingualism as another transitional state.)

I also wouldn't say that languages spread 'much faster' than genes.

A compromise might be that languages take over areas faster than genes. But when it comes to regular minority-group spreading, languages are clearly at most equally fast; and, towards the lower limit, presumably "slower" in the sense that a singular migrant man can leave an identifiable genetic legacy while not leaving no lasting linguistic influence.

[jmcd]

I suspect .that raising children in non-native languages is being more common since the widespread availability of television unfortunately.

[KathTheDragon]

I suspect .that raising children in non-native languages is being more common since the widespread availability of television unfortunately.

Something makes me suspect that television wasn't available when the Indo-Europeans were roaming the steppe...

[jal]

Something makes me suspect that television wasn't available when the Indo-Europeans were roaming the steppe...

Well, in about all Indo-European languages, the respective words for television are cognates. So I wouldn't so easily dismiss that idea!


JAL

Sorry, just being silly. I should go to bed.

[Pole, the]

I suspect .that raising children in non-native languages is being more common since the widespread availability of television unfortunately.

Something makes me suspect that television wasn't available when the Indo-Europeans were roaming the steppe...

YOU CANNOT PROVE THAT