Altaic languages
Altaic is a language family which includes 60 languages spoken by about 250 million people, mostly in and around Central Asia and Far East. The relationships among these languages remain a matter of debate among historical linguists, and the existence of Altaic as a family is rejected by many. Some scholars consider the obvious similarity between these languages as genetically inherited, others propose the idea of the Sprachbund.
Its proponents traditionally considered it to include the Turkic languages, the Mongolian languages, the Tungusic languages (or Manchu-Tungus); to these, most modern proponents add Japanese and Korean. Ainu has occasionally been suggested as a member of Altaic, but this theory enjoys much less support.
History of the Altaic theory
The Altaic family, under the name "Tatar", was first postulated by Schott in 1849, as a family uniting Turkic, Mongolian, and Tungus; he used the name "Altaic" to refer to what would now be called Ural-Altaic (a hypothesis generally rejected.) Castrén (1862) put forward a similar view, but classified Turkic with what we would now call Uralic. Anton Boller suggested adding Korean and Japanese in 1857; for Korean, G. J. Ramstedt and E. D. Polivanov put forward more etymologies in the 1920's. Japonic has commonly been linked to Korean (eg Samuel Martin 1966), and in 1971 Roy Miller suggested relating it to both Korean and Altaic. His suggestion has been taken up and developed by various historical linguists such as Sergei Starostin.
There were some attempts to extend the Altaic family borders by including Ainu (for instance by Street (1962) and Patrie (1982). In recent years it has more commonly been linked to the Austronesian languages, if anything), Tamili, Nivkh or Hungarian languages, but they were rejected by the majority of the scholars.
One of the puzzles of Altaic languages is the nature of the phonetic coincidence r/l - š/s (z), so-called or rhotacism.
Controversy
There are two main schools of thought about the Altaic theory. One is that the proposed constituent language families (Turkic, Mongolic, and Tungusic in the basic theory; with the addition of Korean and Japanese in extended versions) are genetically or 'divergently' related by descent from a common ancestor, 'Proto-Altaic'. The other school rejects this theory (so it is often called the 'Anti-Altaic' school) and argues that the member languages are related by convergence (mainly loan influence).
The Altaic theory is claimed by its opponents to mainly be based on typological similiarities, such as vowel harmony, lack of grammatical gender, an agglutinative typology, and loanwords. In fact, its proponents have put together a large variety of grammatic, lexical and syntactic regular correspondences between the sub-groups of Altaic (eg Ramstedt, Poppe, Martin, Starostin). However, its opponents explain these as loanwords or mutual influence, arguing that, although Turkic, Mongolian, and Tungusic families do have similarities, they are the result of intensive borrowing and long contact among speakers.
The Altaic theory is supported by many linguists, but many other linguists (eg Doerfer 1963) do not regard Altaic as a valid group, and see it as three (or more) separate language families. Other linguists, such as Bernard Comrie (1992, 2003) argue that Altaic may be part of a larger grouping, such as Nostratic or Eurasiatic. In contrast, J. Marshall Ungar (1990) believes that languages such as Korean and Japanese may be part of a macro-Tungusic family.