|Typology of Morphosyntactic Variations of Ergative Constructions in the Batsbi and the Georgian Languages
Languages can be divided into three major types, depending on the extent to which the three main dimensions of relational structuring (semantic roles, information flow, and deistic anchoring) are grammaticalised: “pivotless” languages, with no or little grammaticalization of any of these dimensions,“pure” languages, strongly grammaticalizing only one of them, and “mixed” languages, strongly grammaticalizing more then one dimension.
According to this typology, Georgian is a mixed language, more precisely a role-deixis-oriented language, but the same qualification would apply to Batsbi (or Tsova-tush, the designation now preferably used by native speakers), where ergative case is used provided that the noun is (i) Ag, (ii) I/II person, and (iii) acts according to his/her ‘free will’ and controls the action . Due to this peculiarity, it seems that Tsova-tush developed a fluid-S pattern  from an original ergative pattern which is preserved in other Nakh languages. Although, verb concord (class markers) which represents the category of ‘transitivity’ keeps the original ergative construction of the Nakh languages: It is not restricted to I/II person and the Patient doesn’t show the I/II versus III dichotomy either (no person markers for it). Therefore, in Tsova-tush, category of ‘Transitivity’ is stronger and stands over the category of ‘Communicative Act’; so, the hierarchy T>CA takes place.
Peculiarities of ergative constructions in Tsova-tush, presumably, might be resulted from the Georgian language influence (the Tsova-tush language is spread in Axmeta region of Georgia), as far as the Georgian, strives for formal markedness of I/II:III opposition and grammaticalization of the dichotomy is a dominant in the system of its grammatical categories . But in Georgian, on the other hand, a different, opposite hierarchy should be used to arrive at adequate linguistic structures: CA>T.
Although Tsova-tush is similar to Georgian in one respect (both can be qualified as role-deixis-oriented languages), it also differs from it in a significant way: Because of contact with the Georgian language Tsova-tush distinguishes the I/II versus III dichotomy, although formalization of it takes place only after grammaticalization of class category which is the basis for the prototypical ergative constructions. That is, it keeps original category as more strong and puts it on a higher position in hierarchy.
Such a theoretical approach gives us an argument to conclude that for purposes of a comprehensive description of languages, it is not enough to define languages as merely mixed systems, but also state of which hierarchies they make use.
The hierarchies are defined according to the priority given to marked categories during surface realizations: They reflect dynamic synchronic processes of linguistic structuring. The hierarchies can explain diachronic linguistic changes as well and, we suppose, that they can give us new understandings about the nature of linguistic change in situation of language contacts: When new categories or structures rise in languages because of contacts (or because of language internally motivated variations), the old ones doesn’t disappear and take a high position in a hieratically organized linguistic system.
 Asatiani, R. 1999. Markedness and the Dominant Category. In the Proceedings of the 3rd and 4th Symposium on Language, logic and Computation. Amsterdam: ILLC Publications.
 Braithweithe, K. 1973. Case Shift and Verb Concord in Georgian. Phd thesis. Department of Linguistics, the University of Texas at Austin.
 Dixon, R. M. W. 1994. Ergativity. Cambridge: Cambridge University Press.
 Holisky, D. A.1987.The Case of the Intransitive Subject in Tsova-tush (Batsbi). Lingua, 71:103-132.
 Kibrik, A. 1997. Beyond Subject and Object: Toward a Comprehensive Relational Typology. Linguistic Typology 1, p. 279-346. Berlin-New York: Mouton de Gruyter