The etymon is, of course a Wanderwort (cf. Kartv. *ḳaṭu, late Latin cattus, Slav. *kotъ, Osset. gädɨ, see Abayev 1958, 510, etc.). Klimov (1967, 380-381) tried to explain its presence in the Caucasian languages as a loan from Armenian; but the word no doubt existed already at least in PWC and in sub-protolanguages like PL and PA which makes this theory impossible.

    Even if the word is borrowed from some external source, the borrowing occurred so early that we think it justified to include the word into the common NC wordstock.

    The principal Proto-Av.-And. form must be reconstructed as *gitu. It was preserved only in Southern Av. dialects (Chad. gétu, Gid. geto), being replaced in Northern Av. by the assimilated form kéto (kétu). The Akhv. (Tseg.) form geto may also reflect this archaic form, or (which is more probable) is borrowed from Southern Avar.

    The literary Av. form kéto was borrowed into some of the Andian languages: cf. Akhv. keto, Ratl. keto, Tlan. ketu, Kar. Tok. ketu, Bagv. kitʷ.

    The proper Proto-Andian form is reconstructed as *gidu (which is a result of assimilation from the original *gitu) and is reflected as such in the majority of languages. It was then borrowed in some Tsezian languages: Bezht. gedo, Gunz. gedu (the proper PTs form is *ḳiṭu q.v.).

    The only evident exception is the Cham. (U.-Gakv., Gad.) form č̣eṭa ( < *ḳiṭa) - since all other Av.-And. languages do not show any trace of glottalisation in this root, it must be borrowed, most probably from Tsezian.

    The final *-j is clearly seen in forms like Tab. Kand. plur. gatuj-ir. The original disyllabic structure of the root is evident from the Lezg., Kryz. and Bud. forms, where the dental *t was palatalized before a front vowel (*-tij > Lezg. -ci, Kryz., Bud. -či). Afterwards the final -i in these languages (also in Rut.) was reinterpreted as the oblique stem vowel and was lost in the direct base.

    The forms reveal a variation of reflexes (*a or *ä in the first syllable, *u or *i - in the second). The most probable solution is to reconstruct an original protoform like *k:ätuj (with different directions of vowel assimilation in later reflexes); thus the most archaic is the modern Ag. (Bursh.) form getʷ. It is interesting to note that the root even has traces of old Ablaut *ä/*i, cf. Rut. gät, obl. git-ɨrɨ-.

    See Бокарев 1961, 62; Лексика 1971, 153; Гигинейшвили 1977, 79-80.

    a) *ḱ:aṭ́ǝwǝ > *ḱ:aṭǝwǝ ( > PAK *k:at:ǝ́wǝ, Ub. ǵaṭǝ́);

    b) *ḱ:at́ǝwǝ > *ḱ:ʷat́ʷǝ (with regressive labialisation) > PAT *gʷǝćǝ > *ćǝgʷǝ. [Abdokov (1983, 123) compares the Abkh. form with heterogeneous diminutives: Chech. cic-ig, Av. ic:i-ko 'cat, kitty'. This is certainly wrong because of the obviously different morphological structure of EC forms.]

    It is interesting to note that these two forms exactly match two variants reconstructed for PEC (*gǟṭuj and *gǟtuj).