Comments: A demonstrative stem, primarily adverbial with far deixis, but also used in some other functions. In AA cf. Abkh. a-ná 'there', a-ni (*a-nǝ-jǝ) 'that' etc.; also used as a preverb na- 'there' (directive). It is possible that the 3d person pronominal prefix n(a)- (in some verbal forms) in Abkh. and Abaz. also goes back to the same root.
In AK the stem has survived only as a preverb *na- (Ad., Kab. na-) with exactly the same functions as in AA. Finally, in Ubykh the only traces of this morpheme are the pronominal prefixes of the 3d person n- (sing.), nā- (plur.), obviously related to the corresponding AA prefix *n(a)-.
Comments: PAK *nǝbʁʷV́. Ub. nabʁ́á (in Dirr's notation) is genuinely related to PAK, whereas the later form recorded by Vogt - nǝbʁá - is probably a later loan from Ad. See Abdokov 1973, 61, Shagirov 1, 284.
Comments: PAT *pasa; PAK *pása. The explanation of the Adygh form as 'first, front' + 'to sow' (Shagirov 2, 8) is most probably a folk-etymology, and there is no need to assume that the PAT form (corresponding to PAK quite regularly) is an Adygh loan.
Comments: PAT *pašʷǝ (cf. also Bz. a-páš̌). The meaning 'in front' is probably conditioned by analogy with PAT *pǝ 'nose, front' (or is this really a compound with this morpheme in the first component?). A similar root is present in Kartvelian (PK *pasw- 'root', see Klimov 187), so a loan from Kartvelian is possible (however, the reverse can also not be excluded).
Comments: PAT *ǯ́ǝ; PAK *pč́:a-jǝ (with a usual "tree"-suffix *-jǝ), cf. also Shaps. pč́ajǝ 'beech-tree'. The PAT form can not be related to PAK *č:ǝ-ɣǝ 'tree' (see under PWC *č́:ʷǝ); on the other hand, the semantic shift 'oak-tree' - 'plane-tree', 'beech-tree' is rather common (it happened, e.g., in Greek), and the comparison of PAK *pč́:a-jǝ́ with PAT *ǯ́ǝ 'oak-tree' seems quite plausible.
Comments: PAT *ǯ́ǝma; PAK *pč́:anǝ́. In spite of irregular -n- in PAK (probably result of dissimilation), this comparison (proposed by Rogava 1956, 75) seems reliable enough and better than the comparison of PAT *ǯ́ǝma with PAK *č:́amǝ́ 'cow' (see Shagirov 1, 85, 192).
Comments: PAT *ć̣a-rV 'sharp' (cf. also Bzyb. á-ć̣ar); the same root (without the suffix) is present in PAT *pǝ-nǝ-ć̣a 'nose' (q.v.) - a compound, except for the linking morpheme *-nǝ-, identical to Ub. fa-ć̣á and PAK *pá-ṗc̣a (the first component see under *pʷV).
The PAT *pǝnǝć̣a 'nose' is compared (by Shagirov 1972, Abdokov 1973) with Kab. panc̣ǝv 'tip of the nose'. Absence of this form in Ad., however, speaks rather in favour of this form's being borrowed from Abaz. (Ashkh.) pǝnc̣ǝžʷ 'nostril(s)' - which can be probably analyzed as pǝnc̣a 'nose' + -žʷ 'expressive suffix'. This seems to us more probable than Shagirov's (1977) analysis of pǝnc̣ǝv as pǝ 'nose' + c̣ǝv 'beetle' (?). The WC compound form *pʷV(nǝ)-ć̣a spread also to the neighbour languages: Georg. ṗinčwi, Arm. ṗinǯ / ṗinč 'nostril', Osset. fɨnʒ / fɨnʒ(ä) 'nose, tip' (see Abayev 1958; Abayev lists also Megr. ṗiǯi 'mouth' - which, according to Klimov 1963, does not belong here and descends from PK *ṗir- 'edge'. On the contrary, Klimov (1963) considers the Armenian form to be a megrelism which is hardly possible; he does not list the Georgian form).
Comments: PAK *ṗc̣a. The WC etymology of Shagirov (2,50-51: comparison with Ub. baćǝ- 'to become thick') seems dubious for both semantic and phonetic reasons. Since the EC counterpart of the root means both 'sour' and 'raw' (a natural and frequent semantic correlation), it seems possible to compare also (with a secondary loss of initial labial) PAK *c̣ǝ-ná (Ad., Kab. c̣ǝna) 'raw' (also borrowed in Ub.: c̣ǝna id., see Shagirov 2, 126).
Comments: PAT *-ṗǝ is used as a common suffixed root denoting present tense of static verbs and future tense of dynamic verbs. Glottalisation is here probably secondary - because of standard reduction of final -ǝ and neutralisation of laryngeal features in final position.
Ub. -tʷ(ǝ)- is a general verbal root ('to be, to be there') used with all kinds of preverbs.
In AK we would expect *-t- to correspond to Ub. tʷ; it seems probable that the two general PWC verbal roots, *tV and *ṕǝ, merged in PAK *tǝ- (see *tV).
Comments: PAK *p:c:ǝ. The root has nothing to do with *c:a 'tooth' (despite Yakovlev 1948, 156-157,274,293). PAK *c: can go back to PWC *c:, *ć: or *č:.
Comments: PAK *pč́árǝ. Final *-rǝ may be a suffix (cf. an analogical suffix in the PA form), but the root *pč́a- has no parallels in other WC languages: the comparison with PAT *pǝśǝ-la 'fat' (see Shakryl 1968, 29; Abdokov 1973, 16, 21, 27; Shagirov 2, 29) can not be accepted for phonetic reasons (the correspondence PAK *č́ : PAT *ś does not exist).
Comments: Ub. a-ʁʷǝ-pč́á-n. All forms go back to the PWC compound *ʁʷa-pǝč́:a (with *ʁʷa- "yellow" q.v.). Correspondences are quite regular, thus there is no need to consider the PAT form as an Adygh loan.
Comments: PAK *pčaχʷá. The word presents several difficulties (see Шагиров 2, 27). It was borrowed into Abaz. in the form pχačaqʷa - however, Shagirov is probably right in thinking that the AK word had been reinterpreted in Abaz. as pχača 'coarse-grained' (also an AK loan) + qʷa 'ashes'. Circassian pχa(r)šāχʷa 'sand' is in this case a back-loan from Abaz. Ubykh pšāχʷa 'sand' is also most probably a recent Ad. loanword. Further etymology of PAK *pčaχʷá is not clear.
Comments: PAK *ṗc̣a-ná. The root is also used in compounds *ʎa-ṗc̣á 'bare-footed' (Kab. ʎāmc̣a / ʎāṗc̣a, Ad. ʎāṗc̣a) = Ub. ʎa-ṗc̣ǝ́ and *śħa-ṗc̣á 'bare-headed' (Kab. śħamc̣a / śħaṗc̣a, Ad. śħāṗc̣a) = Ub. ša-ṗc̣ǝ́.
Shagirov (2,49-50) proposes to compare PAK *ṗc̣a rather with Ub. ṗć̣a 'clean' (thus considering Ub. -ṗc̣ǝ a loan from Ad.). This seems to us less likely, because Ub. ṗć̣a has a different etymology (see PWC *ṗǝźV 'clean'), and we prefer to follow Abdokov (1973, 57, 65).
Comments: PAK *ṗč̣́a-. The root is unknown outside the AK area (Klimov's comparison with Abkh. a-bc̣a-ra 'to mow' /see Klimov 1967, 299/ should be abandoned for phonetic reasons). Other etymological attempts (see Shagirov 2, 53 with literature) had also been unsuccessful.
