On the Homeland of the Uralic Language Family
Per Urales ad Orientem. Iter polyphonicum multilingue. Festskrift tillägnad Juha Janhunen på hans sextioårsdag den 12 februari 2012. Suomalais-Ugrilaisen Seuran Toimituksia = Mémoires de la Société Finno-Ougrienne 264. Pp. 91-101. Helsinki, 2012.
Early Contacts between Uralic and Yukaghir
Finnish linguist Jaako Hakkinen based in at Helsinki University has been making interesting and informative comments on GeoCurrents in conjunction with the discussion around the homeland of Indo-Europeans and the dates of Indo-European divergence. He has also made available on his website a number of articles and handouts in Finnish and English on various aspects of Uralic prehistory. I chose one to comment on the current state of our knowledge of the Uralic homeland from linguistic data.
Although the title of the paper refers to the nitty-gritty details of Uralo-Yukaghir lexical contacts, a bulk of the paper is devoted to a discussion of the reasons behind Häkkinen’s choice of the Lower Kama as Uralic homeland (see circle on the map below, from Hakkinen’s Finnish article “Kantauralin ajoitus ja paikannus: perustelut puntarissa,” Suomalais-Ugrilaisen Seuran Aikakauskirja 92, 2009, 56).
Here are some of the highlights from Hakkinen’s research that caught my eye.
1. Yukaghir is not part of Uralic (contra the widely-spread opinion derived from Irina Nikoaleva’s Russian dissertation (Николаева И. А. Проблема урало-юкагирских генетических связей. Автореф. дис. … канд. филол. наук. М., 1988.). Instead, all the lexical parallels (56 by Hakkinen’s count) between the two languages are loans from Uralic into Yukaghir. Interestingly, the relative chronology of these loans as seen in the light of the relative chronology of Altaic to Yukaghir loans require the postulation of two waves of Uralic-Yukaghir contacts, one at the level of Pre-Proto-Uralic, the other on the Samoyedic level.
2. Samoyedic cannot be considered the most divergent among Uralic branches. Hakkinen questions the reliability of language splits derived from lexical innovations. These patterns can result in what he calls “false divergence.” Since phonologically Samoyedic shares many innovations with Ugric, it is unlikely that it split from proto-Uralic earlier than Ugric. This results in a symmetrical west-east pattern of divergence between Uralic languages (see below)
I am very sympathetic to the phonology-only filter for language phylogenies, as “basic vocabulary” arguments and calculations of divergence times on their basis are always unconvincing because words, no matter how “basic” are subject to various forces of change (borrowing, etc.) and deliberate modification by speakers that have nothing to do with population divergence.
3. With the Yukaghir-Uralic genetic link severed and with Samoyedic rolled under “East Uralic,” the conservatism of Finnic and Saami comes to the fore as the primary reason for Hakkinen to seek the Uralic homeland west of the Urals. He is hesitant about shifting the homeland too close to the current area of the distribution of Finnic and Saami because a) Saami place-names attest to the movement of Saami from the Ladoga Lake area within the past 2000 years; b) because there seems to be a foreign substratum in West Uralic pointing to the fact that westernmost territories were at some point occupied by non-Uralic speakers; and c) because there are no borrowed items between proto-Uralic and Early Proto-Indo-European (as there are Late Proto-Indo-European borrowings into proto-Uralic), which one would expect if the Uralic homeland had a more westerly location. The absence of Early Proto-Indo-European loans in proto-Uralic is a key reason for Hakkinen to seek Uralic homeland as far away from the Pontic steppe as possible, provided that the formal conservatism of Finnish and Saami is given credit.
With some help from archaeology, Hakkinen identifies proto-Uralic with Eneolithic Garino-Bor (Turbin) culture 3,000-2,500 YBP located in the Lower Kama Basin.
4. But when Hakkinen extends beyond proto-Uralic, he is is forced to acknowledge, following his mentor, Juha Janhunen, that the structural parallels between Uralic and Altaic are so strong that these two language families had to be in close and prolonged contact with each other at some point.
“Juha Janhunen has repeatedly argued that the Ural-Altaic typological complex is an areally distinct unit with clear-cut boundaries in every direction against languages of different typology, and that the Ural-Altaic typology must have had one original centre of expansion. As the earliest protolanguages of the Al- taic language families (Proto-Turkic, Proto-Mongolic and Proto-Tungusic) can be traced back to Greater Manchuria (up to Mongolia; Janhunen 1996: 216), this view requires that Pre-Proto-Uralic must also have been present somewhere adjacent to them. According to Janhunen there is no significant chance that Pre-Proto-Uralic could have developed a structural typology so similar to the Altaic languages without being in close contact with them.”
So, Hakkinen places Pre-Proto-Uralic in South Siberia in the Sayan Mountains, or again back east of the Urals.
I completely concur with Janhunen and Hakkinen that Uralo-Altaic requires at least a Sprachbund to explain their shared linguistic patterns. A recent global analysis of structural stability between languages reaffirmed the special connection between Uralic, Altaic (Mongolic + Turkic) as well as Indo-European. Kinship studies offers another take on this problem. Most Uralic and Altaic (Tungusic and Turkic) kinship terminologies display a unique and systematic trait called Sliding Generation, or Siberian Generational System (скользящий счет поколений, сибирский генерационный тип, in Russian-language literature). It is not observed in Sino-Tibetan, Paleoasiatic, Indo-European, Nivkh (it may belong with Paleoasiatic, per Michael Fortescue), Japanese, Korean or Dravidian. It was lost from some Southern and Western Turkic languages (Azerbaijani, Nogay, Gagauz, Crimean Tatar, etc.) and borrowed into Ket and, possibly, Yukaghir. However, I do not think that a more northern area as a source for both pre-Proto-Uralic and proto-Altaic can be excluded. It appears that Saami may retain a kin terminological system which is ancestral to Sliding Generation System. It makes a more northerly location closer to the Circumpolar zone a possible source for the Uralic-Altaic Sprachbund. This, in turn, would make perfect sense if the Uralo-Eskimo proposal advanced by Fortescue and Uwe Seefloth turns out to be valid. At this point it is a long-range hypothesis and hence very uncertain. For instance, while Fortescue and Seefloth counted Yukaghir as Uralic, Hakkinen has argued against the genetic nature of the link (see above).
A thorough review!
I have to find out the argumentation of Seefloth. But still, even if Eskimo-Uralic hypothesis is seen as plausible, the “homeland” could still be southern. Native Americans, Yukaghirs, and Yakuts for example are derived from the Southern Siberia. Uralic and Altaic language families point to the original southern area, why not Eskimo-Aleut, too?
According to the comments by Ante Aikio, there may be more common between Eskimo and Samoyedic than other Uralic – a similar case is seen also with Yukaghir and Samoyed at the morphological level.
If you send me your e-mail address, I’ll send you Seefloth’s paper. I’d be curious to know your reaction to it.
Altaic does have a southern distribution (Yakuts and Dolgans aside). Uralic, however, has a northern distribution. You locate it at Lower Kama which is way up north from South Siberia.
I don’t want to open a can of worms with the origin of American Indians. I think a couple of out-of-America scenarios are in fact possible. But for the purposes of this discussion, let’s assume, following mtDNA data, that at least two American Indian lineages (A2, C1) back-migrated across the Bering Strait. Hg A2 back-migration may be related to the Ket-Na-Dene link.
Thank you for the link. The discussants do note a special connection of Eskimo-Aleut with Samoyedic but they also interpret it as a genetic, not a contact link. And the two most intriguing cognate sets connect PU, not just Proto-Samoyedic, with PE. Overall, what strikes me as important is that it’s precisely in the extreme north that we find most conservative branches of Uralic, Finnic and Samoyedic. And this lines up with the Eskimo-Aleut location.
What we may be dealing here with is a signature of expansion from a refugium in extreme North East Asia/Beringia at the end of the Ice Age, which resulted in the southward movement of Altaic, westward movement of Yukaghir and Pre-Proto-Uralic all the way to Scandinavia in several steps and the eastward movement of Eskimo-Aleut all the way to Greenland. In this case, the nexus of similarities between these families may just as likely be due to contact as to common descent.
I would gladly see (and review) the paper of Seefloth! jaakko.hakkinen fat helsinki.fi (clever code to avoid spam-bots.)
How you put it, the east-west-spread sounds quite plausible, especially now when we have Pre-Proto-Uralic in the borders of Central Siberia. I know too little about the American question, so I cannot say anything about that… Except that back-migration sounds very interesting.
Lower Kama is almost directly to the west (50.35′ NLa) from Sayan (50.15′ NLa), and all the Uralic languages north from the Arctic Circle are recent newcomers there.
I e-mailed you the Seefloth paper and the accompanying Georg paper. Let me know if you’ve received them. You do have a point regarding the geographic location of the Lower Kama homeland vis-a-vis the Sayans.
It is a pity that Seefloth didn’t consider any words; it would be a very strong case if both words and morphological features could be presented and the same regular sound correspondences would be seen in them. The ca. 60 basic words I saw for Proto-Eskimo-Aleut didn’t ring any bells from the Uralic point of view…
As Georg commented in the other paper I sent you, Seefloth indeed did not find any lexical parallels between Uralic and EA. He also noted that Afroasiatic shows a parallel situation, where only grammatical markers were used to establish the family unity.
Lexical data is unstable and maybe basic vocabulary experienced drift since the breakdown of Uralo-Siberian. I would be curious to compare Uralic and EA kinship terms. Structurally Uralic (as exemplified by the ancestral pattern preserved in Saami) and EA kinship terminologies are very compatible and I could envision them being closely related. But I’m not an expert in the actual languages, hence it’s hard for me to tell if the actual terms are related.
The Uralo-Eskimo similarity of the verbal structure was already shown by Aurélien Sauvageot. As for myself, I have found some link with Altaic: eskimo imiq = ‘water’/turkish imik ‘fontanel’, eskimo ikiaq ‘between’/turkish iki ‘two’, eskimo kayak ‘boat’/turkish kayïk ‘to glide’.
The last is an old chestnut, but Alonso de la Fuente 2010 disposes of it: kayık < Middle Turkish kaygak id., traceable to Proto-Turkic kad- ‘bend, turn oneself’.
As for qayak, it takes the plural form qaannat in Greenlandish, transparently from qan- ‘come near’ and -yaq ‘locative/instrumental’, showing that qayyak is < internal assimilation. What is more, the Aleut forms iqyax̣, iqax̣, iqyax̣, niqax̣ id. could only be connected to qayaq by a totally ad hoc set of sound changes otherwise unknown, and therefore qa(n)yaq is not even reconstructable to Proto-Eskimo-Aleut, never mind some hypothetical Proto-Altaic-Eskimo-Aleut.
Typological reasoning alone is simply unsound. Hungarian is quite Modern European Sprachbund in typology (as Finnish is not and the other Uralic languages even less), but that does not make it IE, nor does it make Celtic, Armenian, or Indo-Iranian non-IE.
The areal sharing of Uralic and Turkic need not be too much of a problem for a more western Uralic “homeland “. For- the Mongolian homeland of Turkic is nothing but (an admittedly pervasive ) theory. Ie Turkic itself could have had a more westerly early history ( as I think either Sinor
or Golden mention in one of their articles )
When did the pre-proto-Uralics expand into Europe? Anyone know?
The latest linguistic treatment identifies it with the Turbin culture in the lower Kama Basin at 3000 BP. But pre-proto-Uralic is older and is placed in the Sayan mountains in Siberia (http://anthropogenesis.kinshipstudies.org/2012/10/on-the-homeland-of-the-uralic-language-family/)