Uralic–Yukaghir languages

Uralic–Yukaghir
(controversial)
Geographic
distribution:		 Scandinavia, Siberia, Eastern Europe
Linguistic classification: Indo-Uralic ?
Subdivisions:
Glottolog: None

{{{mapalt}}}

The Uralic and Yukaghir languages. After Merritt Ruhlen, A Guide to the World's Languages (1987), p. 64. Color key: light blue: Finno-Permic languages. darkgreen: Ugric languages. yellow: Samoyedic languages. magenta: Yukaghir languages.

Uralic–Yukaghir is a proposed language family composed of Uralic and Yukaghir. It is also known as Uralo-Yukaghir.

Uralic is a large and diverse family of languages spoken in northern and eastern Europe and northwestern Siberia. Among the better-known Uralic languages are Finnish, Estonian, and Hungarian.

Yukaghir is a small family of languages spoken in eastern Siberia. It formerly extended over a much wider area (Collinder 1965:30). It consists of two languages, Tundra Yukaghir and Kolyma Yukaghir.

History

The idea that the Uralic and Yukaghir languages are related is not new. It was advanced in passing by Holger Pedersen in his 1931 book Linguistic Science in the Nineteenth Century (p. 338). Pedersen included both languages in his proposed Nostratic language family (ib.).

A genetic relationship between Uralic and Yukaghir was argued for in Jukagirisch und Uralisch ('Yukaghir and Uralic'), a 1940 work by Björn Collinder, one of the leading Uralic specialists of the 20th century (Greenberg 2005:124). Collinder was inspired by a 1907 article by Heikki Paasonen, which pointed to similarities between Uralic and Yukaghir, though it did not identify the two languages as genetically related (Greenberg, ib.).

Uralic–Yukaghir is listed as a language family in A Guide to the World's Languages by Merritt Ruhlen (1987).

Uralic and Yukaghir are listed as a language family, forming a subgroup of a higher-order family called "Eurasiatic", in Joseph Greenberg's Indo-European and Its Closest Relatives (2000–2002). However, Greenberg did not present a discursive argument for this subgroup, presumably leaving it to the reader to work it out from the data on grammar and vocabulary he presented.

The Uralic–Yukaghir family has been accepted by the American Nostraticist Allan Bomhard (2008:176, citing Ruhlen 1987:64–65) but without presenting any argument for it, and Russian linguist Vladimir Napolskikh.

Arguments

Collinder summarizes his case for, not merely a relationship between Uralic and Yukaghir, but for a Uralic–Yukaghir language family, in his Introduction to the Uralic Languages (1965:30):

The features common to Yukagir and Uralic are so numerous and so characteristic that they must be remainders of a primordial unity. The case system of Yukagir is almost identical with that of Northern Samoyed. The imperative of the verbs is formed with the same suffixes as in Southern Samoyed and the most conservative of the Fenno-Ugric languages. The two negative auxiliary verbs of the Uralic languages are also found in Yukagir. There are striking common traits in verb derivation. Most of the pronominal stems are more or less identical. Yukagir has half a hundred words in common with Uralic, in addition to those that may fairly be suspected of being loanwords. This number is not lower than should be expected on the assumption that Yukagir is akin to Uralic. In Yukagir texts one may find sentences of up to a dozen words that consist exclusively or almost exclusively of words that also occur in Uralic. Nothing in the phonologic or morphologic structure of Yukagir contradicts the hypothesis of affinity, and Yukagir agrees well with Uralic as far as the syntax is concerned.

Collinder goes on to compare Uralic with Indo-European (1965:30–34). He begins by saying, "An Indo-Uralic affinity is far from being strictly proved", and emphasizes that he has not been able to find as many "plausible" cognates for Uralic and Indo-European as for Uralic and Yukaghir, at most about twenty. He then goes on to list 15 of these cognates, with reflexes in Uralic and Indo-European languages, and to discuss similarities between Uralic and Indo-European in morphology. He makes no explicit statement on the subgrouping of the language family that must have existed if Uralic, Yukaghir, and Indo-European are all related. While the similarities he attributes to Uralic and Yukaghir, but not to Indo-European, such as the imperative endings, show closer kinship of Uralic and Yukaghir than of Uralic and Indo-European, the similarities he attributes to Uralic and Indo-European, but not to Uralic and Yukaghir, could also be placed in the balance, for example a plural in *-i.

Collinder appears to enumerate approximately as many morphological similarities between Uralic and Indo-European as between Uralic and Yukaghir and at least as many similarities in pronouns (though his treatment of the pronouns is not very comprehensive), so that at bottom the case he makes, if any, for a closer relationship between Uralic and Yukaghir than between Uralic and Indo-European comes down to the fact that Uralic and Yukaghir, in his view, have about two and a half times as many fairly secure lexical cognates as Uralic and Indo-European — a fact that is certainly suggestive but cannot, in the present state of the question, be considered definitive.

Greenberg (2002) lists 115 lexical cognates for Uralic and Indo-European versus 44 lexical cognates for Uralic and Yukaghir. He lists 35 lexical cognates for Indo-European and Yukaghir. Of these cognates, 27 are common to Indo-European, Uralic, and Yukaghir.

Even if one considers Greenberg's listings to be only the result of inspection, as they do not involve the construction of a comprehensive set of sound correspondences, they at least constitute a suggestive preliminary result. (One should note however that a large number of Proto-Uralic lexemes are traditionally explained as loanwords from Proto-Indo-European.)

The figures could also be skewed by the fact that the Indo-European and Uralic languages are much more heavily attested than the Yukaghir ones. However, the impressive figure is the 44 Uralic / Yukaghir cognates versus the 35 Indo-European / Yukaghir ones. This figure does not by any means suggest an overwhelmingly closer relationship of Yukaghir and Uralic than of Yukaghir and Indo-European, especially since the opportunities for mutual influence were much greater between Yukaghir and Uralic than between Yukaghir and Indo-European, both with regard to the transmission of innovations and the conservation of older forms (on the latter, cf. Sweet 1900:120–121).

It therefore appears that an explicit case for a Uralic–Yukaghir language family, as opposed to a Uralic–Yukaghir kinship at some unspecified higher level, has not been made. Collinder (1965:30) has vaguely implied the existence of a Uralic–Yukaghir genetic node, and in this he seems to have been followed by Ruhlen and Greenberg, followed in turn by Bomhard. The actual relations of these languages remain to be scrutinized in detail. No work of synthesis of a definitive nature has appeared on this subject. As in several other cases of this sort (see Ural–Altaic languages), that Uralic and Yukaghir are related has been argued for in some detail, but how they are related remains uncertain.

It is therefore not surprising that, at the present time (2009), the validity of the Uralic–Yukaghir family is controversial. The questions that need to be answered to make a definitive determination in favor of it or against it are:

In answering these questions, a key additional question is:

The last question is a source of controversy, since historical linguists are divided into two opposed camps that have been called "Geneticists" and "Diffusionists" by Merritt Ruhlen (1994:113–114) and are sometimes caricatured as "lumpers and splitters". This conflict will not soon be resolved. However, the fact that, for instance, no-one still believes in the Hamitic family or that the Dené–Yeniseian family has recently attained general acceptance suggests that linguistic science may yet be able to resolve some of these seemingly intractable issues (cf. Vajda 2008).

Häkkinen (2012) rejects a genetic link between Uralic and Yukaghir, arguing that the grammatical systems show too few convincing resemblances, especially the morphology, and proposes instead that many apparent Uralic–Yukaghir cognates are really borrowed from an early stage of Uralic (ca. 3000 BC; he dates Proto-Uralic at ca. 2000 BC) into an early stage of Yukaghir, while Uralic was (according to him) spoken near the Sayan region and Yukaghir near the Upper Lena River and near Lake Baikal. Aikio (2014) agrees that Uralic–Yukaghir is unsupported and implausible and that several apparent cognates are borrowings from Uralic into Yukaghir, though he rejects some of the examples adduced by Häkkinen as spurious or accidental resemblances and puts the date of borrowing much later, arguing that the loanwords he accepts as valid were borrowed from an early stage of Samoyedic (preceding Proto-Samoyedic; thus roughly in the 1st millennium BC) into Yukaghir, in the same general region west of Lake Baikal.

The evidence of the Uralic–Yukaghir negative verb

A more decisive piece of evidence that Uralic and Yukaghir form a valid genetic node comes from the presence of similar negative verb formations in each. In 1933, the pioneering Indo-Europeanist Holger Pedersen published a comparative study of the verb endings in Uralic and Indo-European. According to Pedersen (1933: §8), the Uralic negative verb was formed from a verb *e- meaning 'be' plus the personal endings, which are identical to the personal pronouns. For instance, for the verb *mene 'go', the Proto-Uralic forms would have been: 1s. *e-m mene, 2s. *e-t mene; 1p. *e-me mene, 2p. *e-te mene meaning 'I do not go', 'thou dost not go', etc., literally 'be-me go', 'be-thou go', etc. The Uralic negative particle *ne 'not' would have originally formed part of the expression, as *ne e-m mene 'not be-me go' etc. What made it possible to drop the negative particle was that there was no morphological or syntactic possibility of confusing the expression with the positive verb, which had the form 1s. *mene-m, 2s. *mene-t; 1p. *mene-me, 2p. *mene-te (still preserved without change in Finnish, except for final *-m > -n) 'I go', etc. The dropping of the negative particle is well-attested by such expressions as French pas, personne, rien and Old Norse engi, eigi (ib.).

Pedersen’s argument implies that the Uralic negative verb is very ancient, for three reasons: (1) The Uralic negative verb represents a worn-down formation and is therefore presumably an old one. (2) It is built on an obsolete verb ‘to be’. (3) Most decisively, it is found in all branches of Uralic, including Samoyedic. Therefore, it is at least as ancient as Proto-Uralic, that is to say, some 4,500 or 6,500 years old, according to the two prevailing current estimates.

The Yukaghir negative verb is unlikely to be borrowed from Uralic, because morphological formations are rarely borrowed. Yet no evidence for the Uralic/Yukaghir negative verb formation is found in the Indo-European languages. It thus seems likely that the Uralic/Yukaghir negative verb goes back to a period of primitive Uralic–Yukaghir unity, from which Indo-European was separate. It could nevertheless be imagined that Indo-European once had this negative verb, but had lost it by Proto-Indo-European times, so this evidence is not conclusive, though strongly suggestive.

The principal pronouns in Yukaghir, Uralic, and Indo-European

MeaningYukaghirUralicIndo-European
'I, me' (first person singular) met *mun, *mina 1 *egom (nominative)
*mene (genitive) 2
'you' (second person singular) tet *tun, *tina 1 *tu (nominative) 3
*tewe (genitive)
'this', 'that' (demonstrative pronoun) tiŋ (Tundra), tuŋ (Kolyma) 'this'
taŋ 'that' 		*tä 'this' 4
*to 'that' 5		 *to- 'this, that'
'who?, what?' (interrogative pronoun) kin 'who?' (Kolyma) *ki ~ *ke ~ *ku ~ *ko 'who?, what?'
*ken 'who?' 6		 *kʷi- ~ *kʷe- ~ *kʷo- 'who?, what?'
*kʷi/e/o- + *-ne 'who?, what?' 7

Chief sources for table: Greenberg 2000, Szemerényi 1996. When Yukaghir alone is cited, it indicates the Tundra and Kolyma dialects have the same form. An asterisk indicates reconstructed forms.

Notes to table

  1. Two different sets are reconstructed. The first is seen in e.g. Udmurt mon, ton, the second in e.g. Estonian mina, sina.
  2. E.g. English mine.
  3. E.g. English thou.
  4. E.g. Finnish tä-mä.
  5. E.g. Estonian too.
  6. E.g. Finnish ken.
  7. E.g. Latin quidne.

Bibliography

Works cited

Further reading

See also

External links

This article is issued from Wikipedia - version of the Saturday, January 09, 2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.