Editing Ural-Altaic languages (section)

==Relationship between Uralic and Altaic==

The Altaic language family was generally accepted by linguists from the late 19th century up to the 1960s, but since then has been in dispute. For simplicity's sake, the following discussion assumes the validity of the Altaic language family.

Two senses should be distinguished in which Uralic and Altaic might be related.

#Do Uralic and Altaic have a demonstrable [[genetic relationship (linguistics)|genetic relationship]]?
#If they do have a demonstrable genetic relationship, do they form a valid linguistic [[taxon]]? For example, [[Germanic languages|Germanic]] and [[Iranian languages|Iranian]] have a genetic relationship via [[Proto-Indo-European language|Proto-Indo-European]], but they do not form a valid taxon within the Indo-European language family, whereas in contrast Iranian and [[Indo-Aryan languages|Indo-Aryan]] do via [[Proto-Indo-Iranian language|Indo-Iranian]], a daughter language of Proto-Indo-European that subsequently calved into Indo-Aryan and Iranian. 
In other words, showing a genetic relationship does not suffice to establish a language family, such as the proposed Ural–Altaic family; it is also necessary to consider whether other languages from outside the proposed family might not be at least as closely related to the languages in that family as the latter are to each other. This distinction is often overlooked but is fundamental to the genetic classification of languages.<ref name=Greenberg2005>Greenberg 2005</ref> Some linguists indeed maintain that Uralic and Altaic are related through a larger family, such as [[Eurasiatic languages|Eurasiatic]] or [[Nostratic languages|Nostratic]], within which Uralic and Altaic are no more closely related to each other than either is to any other member of the proposed family, for instance than Uralic or Altaic is to Indo-European (for example [[Joseph Greenberg|Greenberg]]).<ref name=Greenberg20017>Greenberg 2000:17</ref>

===Shared vocabulary===
To demonstrate the existence of a language family, it is necessary to find [[cognate]] words that trace back to a common proto-language. Shared vocabulary alone does not show a relationship, as it may be loaned from one language to another or through the language of a third party.

There are shared words between, for example, Turkic and Ugric languages, or Tungusic and Samoyedic languages, which are explainable by borrowing. However, it has been difficult to find Ural–Altaic words shared across all involved language families. Such words should be found in all branches of the Uralic and Altaic trees and should follow regular sound changes from the proto-language to known modern languages, and regular sound changes from Proto-Ural–Altaic to give [[Proto-Uralic]] and Proto-Altaic words should be found to demonstrate the existence of a Ural–Altaic vocabulary. Instead, candidates for Ural–Altaic cognate sets can typically be supported by only one of the Altaic subfamilies.{{sfn|Sinor|1988|p=736}} In contrast, about 200 Proto-Uralic word roots are known and universally accepted, and for the proto-languages of the Altaic subfamilies and the larger main groups of Uralic, on the order of 1000–2000 words can be recovered.

Some{{Who|date=July 2010}} linguists point out strong similarities in the personal pronouns of Uralic and Altaic languages, although the similarities also exist with the Indo-European pronouns as well.

The basic [[numeral (linguistics)|numerals]], unlike those among the [[Indo-European languages]] (compare [[Proto-Indo-European numerals]]), are particularly divergent between all three core Altaic families and Uralic, and to a lesser extent even within Uralic.{{sfn|Sinor|1988|pp=710–711}}
{| class="wikitable"
|-
!rowspan="2"| Numeral
!colspan="3"| Uralic !! Turkic !! Mongolic !! Tungusic
|-
! [[Finnish language|Finnish]] !! [[Hungarian language|Hungarian]] !! [[Tundra Nenets language|Tundra Nenets]] !! [[Old Turkic]] !! [[Classical Mongolian]] !! [[Proto-Tungusic]]
|-
! 1
| yksi || egy || ӈобˮ (ŋob) || bir || nigen || *emün
|-
! 2
| kaksi || kettő/két || сидя (śiďa) || eki || qoyar || *džör
|-
! 3
| kolme || három || няхарˮ (ńax°r) || üs || ɣurban || *ilam
|-
! 4
| neljä || négy || тет (ťet°) || tört || dörben || *dügün
|-
! 5
| viisi || öt || самляӈг (səmp°ľaŋk°) || baš || tabun || *tuńga
|-
! 6
| kuusi || hat || матˮ (mət°ʔ) || eltı || ǰirɣuɣan || *ńöŋün
|-
! 7
| seitsemän || hét || сиˮив (śīʔw°) || jeti || doluɣan || *nadan
|-
! 8
| kahdeksan || nyolc || сидндет (śid°nťet°) || säkiz || naiman || *džapkun
|-
! 9
| yhdeksän || kilenc ||хасуюˮ (xasuyu")
| toquz || yisün || *xüyägün
|-
! 10
| kymmenen || tíz || юˮ (yūʔ) || on || arban || *džuvan
|}
One alleged Ural-Altaic similarity among this data are the Hungarian (''három'') and Mongolian (''ɣurban'') numerals for '3'. According to Róna-Tas (1983),<ref>{{cite book|first=A.|last=Róna-Tas|year=1983|contribution=De hypothesi Uralo-Altaica|title=Symposium saeculare societatis Fenno-Ugricae|series=Memoires de la Société Finno-Ougrienne|volume=185|pages=235–251}}</ref> elevating this similarity to a hypothesis of common origin would still require several ancillary hypotheses:
* that this Finno-Ugric lexeme, and not the incompatible Samoyedic lexeme, is the original Uralic numeral;
* that this Mongolic lexeme, and not the incompatible Turkic and Tungusic lexemes, is the original Altaic numeral; 
* that the Hungarian form with ''-r-'', and not the ''-l-'' seen in cognates such as in Finnish ''kolme'', is more original;
* that ''-m'' in the Hungarian form is originally a suffix, since ''-bVn'', found also in other Mongolian numerals, is also a suffix and not an original part of the word root;
* that the voiced spirant ''ɣ-'' in Mongolian can correspond to the voiceless stop ''*k-'' in Finno-Ugric (known to be the source of Hungarian ''h-'').

===Sound correspondences===
The following consonant correspondences between Uralic and Altaic are asserted by Poppe (1983):<ref>{{cite book|first=Nicholas|last=Poppe|year=1983|contribution=The Ural-Altaic affinity|title=Symposium saeculare societatis Fenno-Ugricae|series=Memoires de la Société Finno-Ougrienne|volume=185|pages=189–199}}</ref>
* Word-initial bilabial stop: Uralic *p- = Altaic *p- (> Turkic and Mongolic *h-)
* Sibilants: Uralic *s, *š, *ś = Altaic *s
* Nasals: Uralic *n, *ń, *ŋ = Altaic *n, *ń, *ŋ (in Turkic word-initial *n-, *ń- > *j-; in Mongolic *ń(V) > *n(i))
* Liquids: Uralic *-l-, *-r- = Altaic *-l-, *r-<ref group="note">Treated only word-medially.</ref>