Caucasian languages includes 38 languages spoken by about five million people. It comprises several unrelated families of languages, spoken in the Caucasus, Eastern Europe.
Georgian, Svan, Laz, Mingrelian
- Northwest Caucasian languages or Abkhaz-Adygh languages
- Northeast languages or Dagestan languages
Ghodoberi, Karata, Bagvalal, Tindi,Avar, Dido (Tsez), Hinukh (Ginukh), Hunzib, Bezhta, Khvarshi, Darwa, Kajtak, Kubachi, Lak, Aghul, Archi, Budukh, Khinalugh, Kryts, Lezgi, Rutul, Tabassaran, Tsakhur, Udi
- North Central languages or Nakh languages or Vaynakh languages
The last two groups sometimes are called Nakh-Dagestan languages. They have a number of common features in phonetics and grammar, for example, ergative case and sentence structure, but their genetic relationship is not clear. In addition to these peculiarly Caucasian languages, there are linguistic 'islands' of:
- Altaic languages
Udi is an aberrant form of Lezgi and Dido an aberrant form of Avar.
The Caucasus has the largest concentration of ergative languages in Europe. Ergativity is relevant to all the languages of the Caucasus except for Mingrelian, in which the ergative case has been levelled across all subjects of verbs. All of these language families are characterised by an ergative system; also, they tend to be verb-focused, with much information about nouns encoded in the verb.
All Caucasian language families have been linked with various other language families, some with more success than others.
- Basque has been linked with the South Caucasian languages, due to the fact that Basque is the only ergative language remaining in western Europe. However, the application of the ergative case differs between the families, and this hypothesis is not widely accepted.
- A hypothetical Sino-Caucasian language superfamily has been postulated.
- The Indo-European languages have been linked with the Northwest Caucasian languages, apparently quite convincingly, in a superfamily called Pontic.
- The Northwest and Northeast Caucasian families have been linked with each other to form a hypothetical Common North Caucasian language family. This theory is not yet widely accepted; many of the parallels which have been drawn depend on cognates which may, in fact, be loan words.
- Some analyses of Proto-South Caucasian indicate that this protolanguage possessed just one phonemic vowel; this bears an interesting resemblance to some reconstructions of the Proto-Indo-European language, which has also been reconstructed with just one vowel (although this is not widely accepted). Also, the modern Northwest Caucasian languages possess just two phonemic vowels each, which points towards a link in that direction as well.