According to their own legends, they were exiled to the Yenesian River area from where they live now, perhaps around 0 CE, in the face of fierce invaders (perhaps Indo-Europeans or the Turks), probably from the other side of the Sayan Mountains, which together with the Altai Mountains (aka Altay), and Lake Baikal, form the geographic barrier between Siberia to the North and Mongolia to the South. This area has also been proposed as a homeland for the Altaic languages (which make up the Turkic, Mongolian and Tungistic language familes and amy also be a distant ancestors of Korean and Japanese). And, this area has been proposed an an area in which a pre-Uralic protolanguage may have had a formative period.
They, in turn, were followed by more than one later waves of migration. Most recently, since about the 18th century, Russian influence spread east across Siberia. Roughly in the 13th century, the Mongolians had an empire that spanned Siberia but it swiftly collapsed with not much of a trace in the region, and before that from the middle of the 1st millenium until at least about 900 CE, the Turks expanded across much of Siberia. Indo-Europeans had spread as far as the Tarim Basin by about 2000 BCE (which they held until about 600 CE), and the migration east across Siberia was probably underway by 3500 BCE.
At least in Northeastern Europe, the Indo-Europeans were preceded by the Uralic peoples, whose language survives, for example, in Finnish and Hungarian. The Pitted Ware culture (3200 BCE to 2300 BCE in Southern Scandinavia) is sometimes seen as linguistically Uralic, on the grounds that they were a Mesolithic fishing culture rather than a Neolithic hunting one, like their Comb Ceramic neighbors to the east.
The Pit-Comb Ware Culture (aka Comb Cermamic 4200 BCE to 2000 BCE) is almost surely linguistically Uralic. Although, according to James P. Mallory and Douglas Q. Adams, "Pit-Comb Ware Culture", in Encyclopedia of Indo-European Culture,( Fitzroy Dearborn, 1997), pp. 429–30, via Wikipedia: "some toponyms and hydronyms may indicate also a non-Uralic, non-Indo-European language at work in some areas." One can imagine those toponyms and hydronyms as being associated, for example, with a dead language of a lost culture bearing mtDNA halogroup V also found in Iberia and Northeast African Berbers, that was lost when the predecessors of the modern Saami of Northern Sweden and Finland (perhaps associated with Pitted Ware culture) were assimilated into a Uralic culture from which mtDNA N1a made its way on a "great circle" route into Northern Scandinavia from Siberia and before there from Southeast Asia.
The origins of Comb Ceramic culture may be reflected in its pottery, as well as genetically. "It appears that the Comb Ceramic Culture reflects influences from Siberia and distant China." (per Marek Zvelebil, Pitted Ware and related cultures of Neolithic Northern Europe, in P. Bogucki and P.J. Crabtree (eds.), Ancient Europe 8000 BC–AD 1000: Encyclopaedia of the Barbarian World, Vol. I The Mesolithic to Copper Age (c. 8000-2000 B.C.) (2004).) The 4200 BCE-2300 BCE date for Comb Ceramic may be an underestimate. "Calibrated radiocarbon dates for the comb-ware fragments found e.g. in the Karelian isthmus give a total interval of 5600 BCE – 2300 BCE (Geochronometria Vol. 23, pp 93–99, 2004)," and some accounts suggest Comb Ceramic culture was still in place at 2000 BCE.
Where was this culture found?
Finnmark (Norway) in the north, the Kalix River (Sweden) and the Gulf of Bothnia (Finland) in the west and the Vistula River (Poland) in the south. In the east the Comb Ceramic pottery of northwestern Russia merges with a continuum of similar ceramic styles ranging towards the Ural mountains. It would include the Narva culture of Estonia and the Sperrings culture in Finland, among others. They are thought to have been essentially hunter-gatherers, though e.g. the Narva culture in Estonia shows some evidence of agriculture. Some of this region was absorbed by the later Corded Ware horizon.
The Yenesian languages of which Ket is the last surviving member are probably a part of the same language family as the Na-Dene languages, whose most famous member is Apache. The Apache people moved to the American Southwest around 1000 CE from Northwest Canada, where the rest of the Na-Dene live. (The genetic links between the Na-Dene and the Ket are weak, however, possibly as each group of assimilated neighboring outgroups to the point where the outgroups are a more important source of genetic ancestry than the original core of each of the groups).
To their North are the Eskimo-Aleut language speakers, most famously the Inuit, and an indigeneous population found around Vancouver Island intermixed with Na-Dene speaking groups, speaks a language that may also be a part of the Eskimo-Aleut family. The Proto-Eskimo populations reached the fringes of Alaska from East Asia around 500 CE. Quite credible linguistic evidence associates the Eskimo-Aleut languages , the Uralic language and two other Siberian language groups (Yukaghir (with only an 80 speaker moribund linguistic community left sometimes classed as part of Uralic proper), Chukotko-Kamchatkan (the main living indigeneous language of Northeast Siberia) with a pre-Uralic protolanguage, joining them in a group of Uralo-Siberian languages.
Fortescue argues that the Uralo-Siberian proto-language (or a complex of related proto-languages) may have been spoken by Mesolithic hunting and fishing people in south-central Siberia (roughly, from the upper Yenisei river to Lake Baikal) between 8000 and 6000 BC, and that the proto-languages of the derived families may have been carried northward out of this homeland in several successive waves down to about 4000 BC, leaving the Samoyedic branch of Uralic in occupation of the Urheimat thereafter.
In this scenario, Inuit migration to North America coincides roughly with Altaic expansion, as does Yenesian exile to central Siberia.
The emerging sense is that the Na-Dene, in addition to the Inuits (who are the same people as the Thule culture), preceded by the Dorset indigeneous Paleoeskimo culture (that no longer exists and was probably replaced to a great extent after flourishing perhaps 500 BCE to 1500 CE in much of Arctic North America, although a relict population called the "Sadlermiut survived until 1902-1903 at Hudson Bay on Coats, Walrus, and Southampton islands."), and the Saqqaq Paleoeskimo culture (2500 BCE until about 800 BCE) whom ancient DNA links to the modern related to modern Chukchi and Koryak peoples of Siberia, that preceded the Dorest, all made migrations to North America that were subsequent to that of the remainder of the indigeneous American founding population. Ancient DNA samples show (see page 34) mtDNA haplogroup D in Dorset remains and mtDNA haplogroup A in Thule remains. The timing of the Na-Dene arrival is not too certain, but probably many thousands of years after the ca. 14,000 years ago that the founding group of modern humans arrived in the Americas from Beringia, probably sometime between 8,000 years ago and 1,000 years ago.
The possibility of a late arrival of the Na-Dene to North America could also help to explain the fact that the Na-Dene are rich in some haplogroups that are relatively rare in the Americas and almost absent from indigneous people in Latin America. Those Na-Dene enhanced haplogroups could be the haplogroups of the original Na-Dene population that was diluted over time as the Na-Dene admixed with local populations and their genes in turn, introgressed into other geographically adjacent populations that were not linguistically related to the Na-Dene. Thus, Inuits and Na-Dene might be the main sources of Y-DNA Q (xQ-M3)especially Q-M242 and C3b, and mtDNA haplogroup D2a1, while Y-DNA R1a and mtDNA X2a show very similar North American distributions, although not necessarily Na-Dene and Inuit centered and could be a result of either introgression from later migrations (e.g. one of the two wave of Paleoeskimos that had ceased to be present in the pre-Columbian era) or a split in the original source population. Notably, this reading also suggests that the initial founding population of the Americas was probably smaller than a population size that relies on genetic contributions that actually only arrived with the Inuits or the Na-Dene or other later population waves.
Some mtDNA studies have supported this more than one wave theory, with a Na-Dene wave perhaps 6,000-10,000 years ago. Another study based on internal genetic variation likewise weakly suggested a more recent Na-Dene origin than other Amerindian populations.
There is not a consensus on this point from genetic studies although the Inuit and Na-Dene populations do show stronger genetic links to Siberia than other Amerindian populations. For example, the above linked study from 2010 shows that:
[T]here is a clear genetic HLA relatedness between isolated populations close to Beringia: Eskimos, Udegeys, Nivkhs (North East coast of Siberia) and Koryaks and Chukchi from extreme North East Siberia (Fig. 2), and North West American populations: Athabaskan, Alaskan Eskimos (Yupik) and Tlingit. . . . Asian populations which are geographically not close to Beringia (Japanese, Ainu, Manchu, Singapore Chinese, Buyi) do not cluster with North Americans neither in NJ dendrogram (Fig. 2) or correspondence analysis (not shown).
Finally, Lakota-Sioux Amerindians which have inhabited in North United States, are not related with Asians and West Siberians (Fig. 2) but with Meso and South Americans.
- Meso-Americans. Most frequent haplotypes (not shown), relatedness dendrograms (Fig. 2) and correspondence (not shown) do not relate these Amerindians with any Asian population, including North East Siberians. Haplotypes of Meso-Americans are shared with other Amerindians and one of them with Alaskan Eskimo (Yupik): A*02-B*35-*DRB1*0802-DQB1*0402.
- South-Americans. These Amerindian speaking groups are related to other South-American Amerindians and to Meso-American Amerindians (Fig. 2). Most frequent haplotypes are shared with other American Amerindians, but not with Asians.
An overview of indigeneous Native American genetics recaps some of the relevant facts and issues.
This scenario also means that there are extant linguistic counterparts to Siberia to both the Eskimo-Aleut and Na-Dene languages, with only the remaining languages, which Greenberg lumps together as Amerind, despite the existence of linguistic families within this category and a lack of obvious proto-language links to bind them into a language family, and the assumption supported by archaeology and genetics that the populations speaking these languages once formed a single founding population ca. 14,000 years ago that has not encountered significant outside influence until the arrival of Columbus and in North American, Na-Dene, Eskimo-Aleut, and two waves of Paleo-Eskimo encounters, that left any potential linguistic trace.
(Leif Erikson's group from Iceland ca. 1100-1300 CE apparently did not leave any linguistic or genetic trace in North America, although it left some material culture remains).
Efforts to link the Na-Dene language and Yenesian languages to the Caucasian languages or to the Basque language are in my view utterly unsupported. There is no genetic overlap, the geographic and timing of the historical cultures don't match, and the linguistic similarities, if any, are remote. They may be a remote link between Basque and some Caucasian languages, at a time depth of about 5000 years (about the time depth between the most remote of the Indo-European languages), although I wouldn't necessarily consider it to be solidly established.
Efforts to link the Caucasian languages to the Uralic languages likewise fall short, and suggestions that Indo-European languages and the Uralic languages have a common origin are at the very least, ill established, although the two languages would have been spoken by adjacent populations on the Eurasian Steppe for thousands of years.
The Altaic language hypothesis, is very credible for the core of Turkish, Mongolian and Tungistic languages, and plausible when it comes to further links to Korean and Japanese.
The notion that Sino-Tibetan languages are derived from or genetically related at any reasonable time depth to Altaic, Na-Dene-Yenesian, or Uralo-Siberian protolanguages has very weak support, although the timing and geography are somewhat less prohibitive. The balance of the evidence puts the origins of all four linguistic groups roughly in the Northeastern quadrant of Asia (and surely one or more proto-Amerindian languages must at some point have been spoken in that region as well).