Comments: PAT *žaḳ́a, PAK *žaḳá. The Abkh. form is usually considered as a loan from AK, because the word may be analysed in AK as 'tail (*ḳa) of the mouth (*ž́a)'. However, external correspondences for the root as a whole make us suspect that this may be a folk etymology; in the latter case - since the correspondences are fully regular - the root could be reconstructed as such for PWC. Still, Ub. žāḳ́a 'beard' is certainly borrowed from AK (otherwise we should expect ž́- in Ub.).
From WC the word was borrowed into neighbouring languages, cf. Osset. (Dig.) zaḳä,zeḳä 'chin, beard', Megr. bžaḳe 'beard' (for initial b- cf. Abkh. Samurz. (Marr) a-bžáḳ́a). See Abayev 1969, 285.
Comments: PAT *(a)źǝ (cf. Bzyb. - in Marr's notation - á-ź); PAK (reduplicated) *zazǝ. Ub. -ca may be observed in c̣ʷa-cá 'gall' - assimilation from *c̣ʷa-zá (?); note that Ub. zaza 'gall' (recorded by Dirr), as well as Abaz. zaz (coexisting with the genuine az), is an Adygh loanword.
The correspondence PAK, Ub. z : PAT *ź points to PWC *ž.
Meaning:1 to settle down, get filtered 2 to filter, strain
Abkhaz:a-za-rá 1
Abaza:ra-ʒa-ra 2
Adyghe:zǝ-n 2
Kabardian:zǝ-n 2
Ubykh:zǝ- 2
Comments: PAT *źa (cf. also Bzyb. a-źa-rá 'to settle down') / *ra-ʒ́a- (with a causative prefix and a secondary affricatisation; cf. also Abkh. á-ra-ʒa-ra, Bzyb. á-ra-ʒ́a-ra 'to filter, strain'); PAK *zǝ-. See Shagirov 1, 212.
Comments: The root is attested only in an old compound with *ǵʷǝ 'heart'. Cf. also Ab. gʷǝžʷ-ḳ 'anger'. In Abkh. we have a-gʷǝ-pž̌á-ra 'anger, resentment' which is obviously a later contamination of the same stem with Abkh. a-pž̌a-rá 'to tear asunder'.
Comments: PAK *qʷǝ-źǝ (on the first part see under *q́ʷa). Despite Vogt, Ub. mIa-šʷǝ́ for obvious semantic reasons can not be a diminutive from mIa "apple". More probably, the association with the diminutive -šʷǝ is secondary (although it could have been the reason for irregular devoicing *žʷǝ > šʷǝ). An indirect proof that this was the case is Georg. mažalo 'wild apple-tree', which could have been borrowed from a language like Proto-Ubykh (before the devoicing).
Comments: PAT *ʒ́a (cf. also Bzyb. á-ʒ́a). The AK compound *bžǝ-jǝzǝ is treated by Shagirov (1,90) as *bžǝ '*hand' (cf. also PAK *ʡa-bžǝ́-bǝ 'handful', Ad. ʡa-bžǝ-m 'fist') + *jǝ́zǝ 'full'. We consider, however, the direct comparison of PAK *jǝ-zǝ (jǝ- is a usual pronominal prefix) with PAT *ʒ́a and Ub. ʒa 'span' more justified semantically, and quite regular phonetically - thus, the coincidence with *jǝ́zǝ 'full' is certainly secondary.
Comments: PAT *ʒ́amǝʕʷa (cf. Bzyb. a-ʒ́ámʕʷa). Ubykh has a regular delabialisation ʁʷ > ʁ after a labial.
Within PAK it seems probable to relate here the stem *žaʁʷǝ́ (Ad. žaʁʷǝ 'chin', Kab. źaʁʷ 'base of the lower jaw'), although regularly one would expect a form like *ʒa(m)ʁʷǝ or *za(m)ʁʷǝ. The irregularity should be explained by a secondary influence of the PAK root *ža 'mouth' (q.v.).
Comments: PAK *ǯ́ǝ (the morpheme is treated as a "suffix of reverse action", see Shagirov 2, 76, and is used in compounds with other verbal roots); Ub. sǝ́-j-ǯ́ǝ-n (another Ub. compound with the same root is ʎaq̇a-ǯ́(ǝ)-,sǝ-ʎaq̇á-ǯ́ "to go after smb." (ʎaq̇á 'trace').
Comments: PAT *(a)ʒ́V (cf. also Bzyb. a-ʒ́áʒ́a 'dew'; the Abkh. form is reduplicated); Ub. zʷa-ʒǝ́ 'snow' (a compound with zʷa 'sky'), ʒǝ- 'to snow' (a-ʒǝ́-n), mIa-ʒǝ́ 'hoar-frost' (with an unclear component mIa-). The voiceless *-s- in PAK *wa-sá-psǝ (a compound with wa- 'sky' and -psǝ 'water') is conditioned by the influence of another root, PAK *wa-sǝ́ 'snow' ( < PWC *śǝ q.v.).
Comments: Ub. a-z-ʒǝ́-n. The Kab. word is formally analysed as 'to throw (ʒǝ-) together', but in view of the Ubykh evidence it is rather a trace of an independent verbal root. The correspondence Ub. ʒ : PAK *ʒ can point both to PWC *ǯ and *ʒ.
Comments: PAT *ǯ́ǝ-ḳa (with a not quite clear second component: is it the singulative marker *-ḳǝ - *ǯ́ǝ-ḳa < *ǯ́ǝ-ḳǝ 'a grain of salt' ?). PAK *č́ǝ-ʁʷǝ́ (Shagirov 2, 142 must be right in equating the second component with *ʁʷǝ- 'dry, get dry'). It is not quite clear why the affricate in PAK had become unvoiced (perhaps a secondary dissimilation).
Comments: PAK *pVʡá (labialisation of ʡʷ in Ad. is secondary, due to the labialising influence of p-). Voiced bI- in Ub. is irregular (probably *pI- > *ṗI- (under influence of glottalisation) > *bI- (dissimilation)). The correspondence PAK *ʡ : Ub. q̇ points either to PWC *q̇́I or *q́I:.