Comments: PAT *pǝsa (cf. also Bzyb. á-psa / a-psá, perhaps [with rather strange semantics] Abaz. psa-qʷa 'willow'); PAK *psa-jǝ́ (with a usual "tree"-suffix -jǝ). In Ub. cf. also psǝ-s "fir-wood". See Shagirov 2, 15 with literature.
Comments: PAT *pǝśǝ- 'to die', *ʕa-pǝśa- 'to get tired' (Abkh. ā́psa-ra, Abaz. ʕapsa-rá). PAK *pśǝ-. In Ub. - a compound whose second component (-χʷǝ-) is not quite clear.
The semantic developments 'to breathe' > 'get tired' and 'get tired' > 'die' are quite usual (cf. also the meanings of the corresponding root in EC).
Comments: PAK *pśa. There are no known WC parallels (comparison with the Ub. compound bIǝša 'top of a hill' proposed by Shagirov 2,32 is not satisfactory both semantically and phonetically).
Comments: PAT *mǝśa (cf. also Bz. á-mśa). PAK *-psa in compounds: *ʎa-psá 'root, bottom part' (in Kab. 'base') /here *ʎa- is obviously 'foot, bottom'/; in Kab. q:ʷa-ps 'root' the first part is probably *q:ʷa 'horn' q.v. Ub. ʎa-msá (def. á-ʎamsa) is an exact match of PAK *ʎa-psá.
Comments: PAK *psǝ (reflected only in Kab.), *ḳa-psá (a compound with *ḳa- 'end, tail'; besides the Ad. form cf. also Kab. č̣ā-psa id.). See Shagirov 2, 132.
The PWC reconstruction is tentative because PAK *s can go back to several sources: *s, *š or *ś.
Comments: Ub. 3d p. a-pstá-n. Dumézil 1975 considers the Ub. word as a loan from Adygh., deriving the Adygh. verb from psǝ 'water'; Shagirov 2,36 also supports this etymology. It is possible, but seems to us less likely in view of the external evidence.
Since there are no AA parallels, it is impossible to distinguish between reconstructing *-š- or *-s- in the root.
Meaning:1 to graze (at night) 2 to spend a night 3 hut
Abkhaz:a-pχ́a-rá 2
Abaza:pχ́a-ra 2
Adyghe:pš́ǝ-ṗa 3
Kabardian:xa-pśǝ-n 1
Comments: PAT *pǝx́a-; PAK *pš́ǝ-. In PAK the verb has been preserved only (with a locative prefix) in Kab. (Abdokov gives also the meaning 'spend some time', whereas the Kab.-Russ. dictonary only gives the gloss 'to graze (at night, of horses, oxen etc.)'. Adyghe reflects only the compound *pš́ǝ-ṗq̇a 'place of temporary abode', 'hut' (borrowed in Ub. as pš́ǝṗq̇á 'lieu d'assemblée en plein air'); cf. also another derivate in Kab. pśǝ-ʡa 'hut'.
See Abdokov 1983, 128-129 with the correct AT-AK comparison (the author quotes also an Ub. verb pš́ǝ- 'to abide, stay temporarily' - which we were unable to identify and which is probably faulty: in Ub. we should regularly expect a form like pśǝ-). Because of the obvious AT cognate, Shagirov's identification of the present root with PWC *pǝʎ:ʷa 'to blow' (q.v.) should certainly be rejected.
Comments: PAT *pǝħa (with a secondary dissimilative delabialisation); PAK *pχʷǝ. The root must have originally meant 'woman' in general and has preserved this meaning in PWC *pǝχ́ʷA-śʷǝ: PAT *pǝħʷǝśǝ 'woman' (Abkh. a-pħʷǝ́s, Bz. a-pħʷǝ́ś, Abaz. pħʷǝs) = Ub. pχ́ašʷ 'wife'. Cf. also Ub. pχ́a-ʒá 'wife, woman'.
Comments: The PAK form *pχǝ corresponds very well to Abkh. and Ub. (semantically: *'root' > a) '(maize) stalk'; b) 'carrot'). The comparison with Abaz. šʷχǝ 'carrot' (Shagirov 2,24) is phonetically impossible. Adygh > Osset. äpχä 'carrot' (see Abayev 1, 171).
Variations in labialisation (*χʷ / *χ) are explained by dissimilation to the initial labial consonant.
Comments: PAT *pǝza (cf. also Bzyb. a-pǝ́za); PAK *cʷa-bʒá.
The first part of the Adygh and Ub. compound, despite Shagirov 1977, 104-105, is connected neither with PWC *c:ʷǝ 'ox' (q.v.) nor with PWC *zʷa 'to plough' (q.v.), but reflects a separate PWC root (see under *Hrājc_ū). Again, pace Shagirov ibid., since the correspondence between Abkh. and Ad. here is quite obvious, it is impossible to connect the second part of the compound with PAK *bzǝ 'to cut' (q.v.). Finally, since Ub. has a palatal ź, the Ub. form (again despite Shagirov ibid.) can not be regarded as an Adygh loanword. The correspondence between PAK *-bʒa, PAT *pǝza and Ub. -bźa is quite regular.
Comments: PAT *mǝza (cf. also Bz. á-mza 'candle, splinter'; -čǝ- in Abkh. goes back to a-čá "wax"). PAK *bzǝ́-jǝ. Ub. def. á-mʒa.
Shagirov (1, 94 - 95) relates PAK *bzǝ́jǝ to *bza "tongue"; the two roots may indeed have contaminated within some compounds, but originally have been, of course, quite different. The form *bzǝ́jǝ is also used in Kab. to denote "maize stalk" and both in Ad. and Kab. within the form Ad. q:ā-mzǝj, Kab. q:ā-bzij "bird's feather" which Shagirov (1, 679) plausibly explains as a distortion of q:az-bzǝj(ǝ) "goose feather" (q:az < Turk. qaz 'goose').
Comments: The basic meaning of the root *(Pǝ-)ʒ́ʷǝ is "dusk, evening (or morning) dawn". In Abkh. it was lost, and in Abaz. stayed only within a compound with (a)la- 'light' (same as in la-šara 'light', historically = "eye").
In PAK the root *pćV ( < *P-ʒ́ʷǝ with devoicing) is used in compounds *pćǝ-ħa 'evening' (listed above) and *pća-dǝ- 'morning, tomorrow' (Kab. pśa-dej 'tomorrow', pśa-dǝ-ǯǝz 'morning', Ad. pč́a-dǝž́ 'morning'). Finally, in Ub. the root -ǯ́ is also used only in compounds: zʷapsǝ́-ǯ́ 'evening' (with zʷapsǝ́ id.) and sʷǝ-ǯ́ '(morning) dawn' (with sʷǝ id.).
See Shagirov 1,69 (Shagirov is right in thinking that Abaz. baχ is a Kabardian loan - pace Abdokov 1973,40; there is, however, absolutely no reason to consider the Adygh. root an Arabic loan).
Comments: For PWC a compound *bǝ̃lV-p:ǝ̃sa (**'eye-hair') should be reconstructed. The first part is normally reflected in Ub. bLa- and PAK *na-, but in PAT replaced by an obscure component *ǯ́ǝ- (PAT *ǯ́ǝ-mǝsV). In Abkh. a further assimilation has occurred (*ǯ́ǝ-ms > ǯ́ǝ-mš́). The strong *p:- in PWC is suggested by the PAK form *-p:c:a < *-p:sa (PAK *na-p:c:á).
