Uralic (U) and Yukaghir (Yuk.)
The Uralic languages represent a wholly uncontroversial linguistic unit that likely is genetically related to the Yukaghir languages of far north-eastern Siberia (Collinder 1940).
o Yurak (extinct)
o Kamassian (extinct)
o Mator (extinct)
- Hungarian (attested already from the 10th century)
o Finnish & Estonian
o The Saami languages
This traditional branching of the Uralic tree has been challenged, e.g. by Carpelan & Parpola who suggest that Samoyedic only is an early offshoot from the same branching as Proto-Ugric (2007: 135); the consequence is that Proto-Fenno-Ugric qualifies as the de facto Uralic protolanguage, but the following bifurcations appears valid, especially in field of lexicography (cf. Janhunen 1998: 461):
[Fenno-Permic] + [Ugric] = Fenno-Ugric (FU)
[Proto-Samoyedic] + [Fenno-Ugric] = Uralic
Geography and history
The modern Uralic languages are scattered over a vast area from Hungary and Northern Scandinavia in the west to the Taymyr Peninsula in the east. Due to the coincidence of history, the national languages Finnish, Hungarian, and Estonian are the most accessible Uralic languages, and especially the former has been the immediate entry to Uralic studies, and not always undeservingly so (on conservative and innovative features of Finnish compared to other Uralic languages, see Korhonen 1981 and Posti 1954). Attestations only appear with the first sporadic contributions in written form in Hungarian at the end of the first millennium CE, which, coincidentally, is about the same time East Slavic expansion into the Volga region ushered in a millennium of constant Russian encroachment on the Uralic speech communities. The incipient stages of this process can be adduced from the disappearance of the tribes Meschera, Merya, and Muroma, recorded only in the Primary Chronicle of the eastern Slavs. The discontinuous distribution is thus largely attributable to continuous assimilation of the native Uralic tribes, recently continued with the death of Kamassian with its last speaker in 1989. The reconstructed culture was based on a huntergatherer economy, possibly with incipient reindeer herding, since highly specialized vocabulary pertains to this domain (Häkkinen 2001). This substantiates its original point of dispersal in similarly northern regions, cf. also the treatment of the ‘cloudberry’ (§ 4.1, item 95), and, while the exact location is contested, the last point of Uralic unity may reasonably be found within the extremes of the present distribution between 6.000 and 4.000 BCE (Bakró-Nagy forthc.: 13). Alternatively, Nichols suggests that the spread originated somewhere far east of the Ural Mountains to its present distribution (1997: 141), a notion possibly supported by its affinity with the Yukaghir languages of far northeastern Siberia.
Linguistic affinities with PIE
The comparison of the entire PIE and Uralic stock is warranted by great resemblance in very core linguistic material such as pronouns and grammar (for an early and prudent comprehensive treatment of the verbal systems, see Pedersen 1933). A small handful of lexical items have traditionally been added to this list (e.g. water, item 128, and name, item 61), but other features, most prominently the consonant inventories, poorly match in the reconstructed stages.
For the Indo-Uralic hypothesis to be written off completely, a comprehensive explanation of the glaring morphological and pronominal correspondences need be presented, and a period of intimate contact between the speakers of the Uralic and IE proto-languages is thus usually implied in most criticisms of the Indo-Uralic theory, but the systemic transfers are seldom treated in-depth; a rare exception is Wagner (1967, see Bjørn 2016: 12f. for a methodological criticism), treating a marginal enclitic form with very limited bearings on the system as a whole, while Rédei’s complete rejection of the common pronominal traits as “Lautsymbolismus” falls incredibly short of the mark (1986: 19). If the similarities in the pronominal and verbal systems of PIE and Uralic are loan phenomena, it should be expected that more vulnerable parts of the vocabulary be borrowed, too. Uralic adstrate is similarly a part of Trubetzkoy’s theory on the ultimately Caucasian identity of PIE (§ 3.2.3).
Genetic affinity: the Indo-Uralic hypothesis
The lure of Indo-Uralic has drawn scholars to the theory for more than a century and a half, from Thomsen (1869) through Hyllested (2008) and the present author (Bjørn 2016 [Add. now rather Bjørn forthcoming(a)]), to account for the sometimes obvious consistencies, and sometimes the seemingly incompatible discontinuations. Some considerations are thus in order to qualify the objective approach to the present study:
1. The grammatical systems of Indo-European and Uralic are demonstrably transposable, and the phonological systems are not, as has been suggested elsewhere (e.g. Janhunen 1999: 212-215), fundamentally incompatible; it may be noted here in passing that none of the two conceivably oldest branchings of IE requires the level of complexity in articulation of the velar series as is traditionally reconstructed.
2. What remains, however, to establish Indo-Uralic as a viable language family is a shared lexicon to facilitate a thorough phonological analysis; lexical borrowing is common, and several layers of loans between the two established language families have been demonstrated, so it necessarily comes down to the basic vocabulary to provide the core evidence for shared heritage. Few obvious cognates have been identified, perhaps most notably PIE *wed-, Finnish vesi (nom.), veten (gen.) ‘water’ (item 128), and this paucity naturally constitutes a red flag, and despite glottochronological efforts (cf. § 1.5.1), the rate of lexical substitution remains undetermined. Much effort has been invested in identifying loans directly between the proto-languages (cf. Koivulehto 1991, 1994, 2001; Joki 1973; Rédei 1986) which may, indeed, have been a widespread phenomenon.
3. Claims that face-value correspondences are too alike to be of common heritage (Koivulehto 2001:257f.), must, however, be regarded as unsubstantiated. It remains to be established what developments may conceivably separate PIE and the Uralic proto-language, but tentative inquiries into the pronominal systems suggest that the systems may actually be quite alike in some aspects, cf. tentative correspondences (e.g. Bjørn 2016 and Kümmel 2015). The blank dismissal on grounds of similarity alone, then, would disqualify shared heritage of English name and Sanskrit nāman-, Latin nōmen, where the inclusion of Finnish nimi only stands out in a paradigm of which we “know” that it is unrelated.
4. Phonotactic prohibition of initial clusters is internally reconstructable for Uralic, but it need not be so for all posterity of the family. In the wordlist below, if a complex (P)IE onset is compared with a simpler Uralic ditto, it is assumed that the item transferred from the former to the latter. If, however, simplification of initial clusters in Uralic is an internal innovation, the pre-proto-language may have had higher complexity, and the conclusions reached have to be reexamined. This is speculative, and for the direct comparison there is no need to entertain such concoctions; however, in a few instances, e.g. ‘mushroom’ (item 117), distribution suggest that the Uralic form is older.
5. It is of some lexicographic curiosity that of only 18 secure Uralic cognates (Häkkinen 2001, employing the very conservative restrictions of attestation in all branches), Helimski finds half of them in PIE, too (2001: 196 fn. 19); most relevant
for the present discussion are the non-pronominal lexemes Fin. ala– ‘under’ ~ IE *Hel– ‘deep’ (not treated here); Fin. nimi ~ IE *h1neh3m– ‘name’ (item 61); Fin. punoa IE *spen– ‘to spin’ (item 116). He similarly questions the lack of “useful” borrowings, i.e. items with a clearer transitional value and concrete content, in the shared proto-material (2001:199). Laxing the strict criteria for complete in-family representation, more items are added, of course. To consider Samoyedic an offshoot on the Ugric branch allows a wider set of possible cognates; around 700 compared to the mere 130 or so when Samoyedic is treated as the original
bifurcation (Carpelan & Parpola 2001: 77).
6. While the classical PIE e-o-Ø Ablaut system has no parallel in the Uralic languages, an older stratum of common vowel gradation may be visible (Bjørn 2016: 18f.).
7. Carpelan & Parpola suggest that satəmization is an areal phenomenon “possibly triggered by the Proto-Finno-Ugric substratum influence upon the Pre-Proto-Indo-Aryan [of the] Abashevo culture” (2001: 131), cf. Balto-Slavic, Indo-Iranian, Albanian, Armenian, and Uralic. Secondary palatalization is also widespread in Tocharian that nonetheless is a centum branch.
8. Campbell’s (1990) proof of recurrent similarities in the arboreal stock of IE and Uralic is unmistakable. The correspondences need only be explained either as common inheritance or loan- or wander words. It is, indeed, peculiar to find such
a broad spectrum of trees represented in both families. A major caveat, however, is that the IE stock in most cases only is represented by the Western branches, i.e. that reflexes are lacking in Tocharian, Anatolian, and Indo-Iranian. Without neglecting the significance of this dearth of evidence, it is certainly conceivable that the Eastern languages have substituted the vocabulary as they encountered
different kinds of landscape. These items are not etymologized.
9. The chronologies of the archaeological cultures assigned by Carpelan & Parpola (2001) are suggestive (cf. § 1.3.1)
[ADD: It seems likely that Uralic instead spread from the Altai-Sayan region at a later point.]:
From what have been gathered above, a hypothesis that can be tried against further evidence may be constructed as a tentative relative chronology of an Indo-Uralic language family with established events:
1) (Tentative) Common Indo-Uralic proto-language is spoken by hunter-gatherers in the forests of the Volga-Ural region. The stop inventory conceivably medium sized.
2) (Established) The PIE language community experiences great societal change on the Pontic-Caspian steppes.
3) (Tentative) Meanwhile, what later becomes the Uralic language community remains on the northern fringes of the IU dialect continuum and partook in the regional palatalization known from the IE satəm languages.
4) (Tentative) Fundamental substitution of the vocabulary occurs in one of the constituent families (perhaps most likely PIE – from Caucasian)
5) (Established) Anatolian leaves PIE before the development of the Core-IE aspect system; the nature of the stop system is inconclusive.
6) (Established) Tocharian similarly departs at an early stage. The stop system is still inconclusive.
7) (Tentative) The Uralic languages more or less continue the Indo-Uralic culture and gradually diffuse; there are no traces of cultural revolution comparable to what IE underwent.
A shorter schematization with a PIE focus would be:
Indo-Uralic → Caucasian influence defines PIE → Anat. departs → etc.
Anywhere along this line, lexemes may enter the language, either by loans or internal derivation. Truly ancient words without comparanda may also disguise themselves in this category, but are, for the present time, indistinguishable from innovative features. Correlations between Anatolian and Uralic therefore require special attention (cf. Uesson 1970: 96). Later loan relations between IE and Uralic languages, of which there are plenty, are not considered in the present treatment; only, of course, so far as to establish whether a proposed etymology is warranted for the proto-languages and not just individual branches hereof. The relationship between IE and Uralic has thus been investigated at intervals, yet without any significant breakthrough in the sense of a lasting acceptance of the phylogenetic link. A look at PIE through the lense of Caucasian, as has been suggested, may enlighten the endeavor, akin to how Modern English is incomprehensible without the knowledge of sustained superstrate influence of Norman French (cf. Ragot 2011). The attempt at a stratification within this publication will provide a framework for assessing the hypotheses of the suggested Uralic and Caucasian lexical elements in PIE.
Collinder, Björn (1940) Jukagirisch und Uralisch
— (1955) Fenno-Ugric Vocabulary
— (1957) Survey of the Uralic Languages
— (1960) Comparative Grammar of the Uralic Languages
Nikolaeva, Irina (2006) Historical Dictionary of Yukaghir
Rédei, Károly (1988-91) Uralisches Etymologisches Wörterbuch
Sinor, Denis (1988) The Uralic Languages
Addendum: Turkic (or Chuvash-Turkic)
Historically Turkic languages have come to play an important role in the geographies treated in the present paper, but since their presence here demonstrably is connected to a historic spread (cf. Mallory 1989: 147f., Nichols 1997: 132f., and Poppe 1965: 60) the family is not relevant in the discussion of the earliest loan relations of PIE. It will be noted, however, that when the Turkic languages seem to partake in a given set of comparanda, most often in relation to the Uralo-Yukaghir languages, the comparandum is given in the discussion. These instances may be effects of Wanderwort phenomena or ancient genetic affinity, but, as stated, impertinent to the present inquiry.
[Add: See my presentation on how to detect early linguistic interfaces across the steppes: Grain on the Silk Road, Eurasian Corridors presentation (Jena 2019)]