Comments: PAT *pǝśǝ- (cf. also Bz. a-pśǝʒ, Ashkh. psǝz). Both in Abkh. and Abaz. the root is present in compounds with unclear second components (-ʒ in Abkh., -lačʷa in Abaz.).
PAK *p:c:a : the pure root has meant 'big fish'; most common is the diminutive derivate *p:c:a-źájǝ (Ad. pcaźǝj, Kab. bʒaźej).
We have to reconstruct initial *p: to account for the AK progressive assimilation (*p:c:a < *p:sa); both in PAT and Ub. the assimilation had been regressive. Numerous attempts to analise the word on WC ground and to explain it as an IE loan (see Shagirov 1, 82 with literature) seem to us unsuccessful.
Comments: The word is lacking in Ad. (bawa- 'to kiss' is quite a different root, despite Kuipers 14), thus *b- or *p:- may be reconstructed for PAK, and, consequently, for PWC. Despite Shagirov 1,77 the Kab. word is certainly not a Turkic loan, and the Ubykh word - not a loan from Kabardian.
Comments: PAT *pǝsa- (used only with preverbs, cf. also *ḳa-pǝsa- 'to throw, scatter'). See Shagirov 1962,53, Abdokov 1973,72, Shagirov 2,72 (with literature). In view of this etymology, the Ubykh merger with wa-s(ǝ)- 'to sit in' appears secondary.
Comments: PAT *baɣ́a; PAK *paɣá; in Ub. - a metathesis (ʁabǝ́ < *baʁǝ́). Semantically and phonetically the forms are very close, and the correspondence PAT *ɣ́ : PāK *ɣ : Ub. ʁ points to the PWC palatalised lateral *Ł́. Therefore Shagirov's explanation of the Adygh form as *pa 'nose' + *bɣǝ 'mountain' is hardly acceptable (although an analogy with *pa 'nose' is possible).
Comments: PAT *ħa (used with locative preverbs); PAK *pχǝ-. The AT form is usually not compared with Ub. and AK (see Mészáros 1934, 223; Shagirov 2, 24), but the comparison seems quite plausible both phonetically and semantically.
Comments: PAT *pǝ-nǝ-ć̣a (cf. also Bz. a-pǝ́nć̣). Etymology of *-ć̣a (= Ub. -ć̣a) see under *(p-)ć̣a. PAK *pa. The morpheme *pʷA 'nose, front' does not occur independently outside AK, but is present in many compounds both in AA and Ubykh (cf. PAT *pǝ-lA 'meet', *pǝ-cǝ '(front) tooth' etc.; Ub. fa-La 'face' /'nose'+'eye'/ etc.).
The PAT form *pǝnǝć̣a can not be kept apart from Osset. fɨnʒ 'nose' (see Abayev 1958, 497); since the PAT form has a clear morphological structure, we must acknowledge that it is also the ultimate source of Georg. ṗinčwi and Arm. pinǯ 'nostril'.
Comments: PAT *ʕa-pǝ-nǝ (-nǝ is a locative/temporal nominal suffix). The PAT and Ub. forms go back to a PWC compound *ʁ́Iʷa-pʷV (*ʁ́Iʷa- 'time, season' q.v.).
Comments: Both in Kab. and Ub. we have compounds, probably with PWC *bzǝ 'water' (with an irregular reflex in Ub. because of the word's expressive nature).
Comments: PAT *bǝzǝ-jA; PAK *q:a-bzá (Shagirov 1, 218 regards the q:a- as a directive preverb). The root *bza- can also be seen in the compound verb *č:a-bza- 'to get pure, settle (of water)'.
The Ub. form with glottalisation suggests PWC *ṗ-; the PAT reflex *-z- confirms it (if in PWC we had *bǝć̣A, then even after assimilation PAT *-ź- would be expected; on the other hand, *-ź- normally yields PAT *-z-). However, earlier glottalisation must have resided on the affricate, since the initial labial certainly is a historical prefix and there were no glottalised labial prefixes in NC or WC.
Comments: PAK *ṗč̣á-śa (the second component is not quite clear; without it cf. probably Kab. ṗṣ́i-j 'plane seeds' = "belonging to leaves", see Shagirov 2, 53; despite Abdokov 1983, 107 it is hardly a separate WC root). In Ub. cf. also ṗć̣á-wa 'leaf', ṗć̣ā́-d(ǝ)ʁa 'paper'. The correspondence PAK *č̣́: Ub. ć̣ points unambiguously to PWC *ƛ̣.
Comments: PAT *-pǝśa (cf. Bzyb. -pśa), used in compounds: Abkh. án-psa 'step-mother', áb-psa 'step-father', Abaz. an-psa,ab-psa;pa-psa 'step-son'. PAK *-ṗasǝ (also used in compounds: Ad. nǝ-na-ṗʷas, Kab. āna-na-ṗas 'step-mother', Ad. tǝ-na-ṗʷas, Kab. āda-na-ṗas 'step-father'). The PAK form is impossible to separate from PAT *-pǝśa, thus we can not accept Abitov's and Shagirov's treatment of the AK morpheme as compound ("sitting instead of"), see Shagirov 1,61, as well as the derivation of PAT *-pǝśa from the verb *pǝśaxǝ- (Abkh. a-psáχ-ra, Bzyb. a-pśáx-ra, Abaz. psaχ-ra) 'to change' (Shakryl 1968, 37). The latter explanation should be rejected also for phonetic (accentological) reasons. Abdokov (1983, 141) also rejects PAK *-ṗasǝ as a compound, adducing instead Kab. bǝfa ( < *bǝsʷa) in bǝfa-q:ʷa 'step-son', bǝfa-pχʷ 'step-daughter': however, the source of these words is unknown to us.