Notes: Reconstructed for the PEC level. Correspondences are regular. The semantic development must be explained as (*'milk') > 'sour milk' > 'alcoholic beverage made of milk' (a rather common drink for steppe and mountain cultures).
Notes: The comparison between EC and WC (PAK) seems reliable both phonetically and semantically. See Abdokov 1983, 112.
Hurrian in Arraphe had a word *šann-orǝ 'plum-(tree), medlar' ( > Akkad. šannūru, perhaps also Sum. šennur id.), which can be compared with EC *swɨ̆nē. Other dialects, however, had *šall-orǝ ( > Akkad. šallūru, Arm. salor) with a secondary -l- (see Diakonoff-Starostin 1986, 34).
The Caucasian word (with the meaning 'grapes' > 'wine') penetrated into Scythian *sana, Osset. sän / sänä (despite Abayev 1979, 67, the Indo-Iranian origin of the word is highly dubious). Note that usually this root is confused with PNC *ʒ́w[ǝ̆]nʔi 'honey, wine' (q.v.) - etymologically quite different.
Notes: Reconstructed for the PEC level. Correspondences are regular (except that there are usual doubts about the extremely rare vowel *ü). Abdokov (1983, 102) compares PAK *-sa in *ḳaʁʷa-sa 'soot', but the element is rather obscure, and the phonetic side of the comparison is not convincing.
Notes: Reconstructed for the PEC level. The Lak. form is reduplicated (rather a frequent phenomenon in plant names).
In the Darg-Lezg. area there exists a stem denoting 'garlic' - *swälgV - which can be probably traced back to the same root (if we assume that -gV is a suffix here, possibly, deglottalized from diminutive *-ḳV), cf. Darg. Chir. s:urge 'garlic'; PL *s(:)älk: > Lezg. serg (erg. serk:i), Ag. serg, Kryz. sarg 'garlic', Bud. sǝrg 'ramson' (-l- is indicated by the fact of preservation of -r- in Kryz. and Bud.; real *-r- would have been lost in this position).
Notes: An expressive root with metatheses (PD *ḳamš:a < *š:amḳa) and irregular changes; quality of the medial velar consonant is hard to establish.
Although very distorted, the root may belong to the set of early Indo-Iranian loanwords in EC: cf. Osset. sänɨkk 'kid' < Indo-Iran. *sčanika-, cf. Bel. šanikh,šinik 'kid', Yagn. sanča etc. (see Abayev 1979, 71-72).
Notes: Reconstructed for the PEC level. Except for the metathesis, correspondences are quite regular. Cf. also HU: Urart. u/ošm-ašǝ 'force, strength' (see Diakonoff-Starostin 1986, 23).
Notes: The word was widely borrowed from Av. into other languages of Daghestan - probably because in Av. it came to mean 'snuff' (an object of busy trade). However, in Av. and PL the forms are probably genuine (because of phonetic and semantic differences). It is possible that *-ṭV is originally a suffix here (see notes on the Av. form); in that case we can compare Hurr. šūnǝ 'breath, soul, self' (see Diakonoff-Starostin 1986, 32). [Cf. also Georg. c̣winṭli 'snot' > Bacb. sinṭr id.?].
Notes: Cf. also HU forms: Hurr. šeɣ-iri 'alive', šeɣ-ori 'fate' or 'life', Urart. šu/oχ-ori / šeχ-eri 'alive' (see Diakonoff-Starostin 1986, 36).
Both phonetically and semantically the EC-WC parallel seems quite plausible. PWC has a frequent labial prefix. In a position following a labial WC labialized consonants usually are delabialized, but not in this case (PAA *ś : PAK *ś point unambiguously to *śʷ; in Ubykh, in fact, delabialisation had occurred - otherwise we would have pšʷa- - but this is a regular later Ubykh process). Lack of delabialisation can be very well explained if we assume a form like *pǝ-śǝHwV > *pǝ-śǝwV (not *pǝ-śʷV) for the early period, when the after-labial delabialisation process was active. For phonetic reasons it seems better to keep apart WC *pǝsV 'soul' (which is superficially closer to PN *sa), which has a quite different NC etymology (see under *ʡăms_a).
Notes: A Lezg.-WC isogloss. The PWC form can be analyzed as "light inside" (with the common WC morpheme *Łʷa- 'inner, inside; to enter'). In this case the correspondence is established between PL *s:ɨlä- and PWC *-sǝ. It is quite regular and seems probable, but must be approached with some caution because of the scarcity of reflexes in EC languages. See Абдоков 1983, 150.
Notes: Reconstructed for the PEC level. The original meaning was probably 'long (curved) bone', with later developments > 'long bone, spine'; 'curve, angle, corner' (cf. a combination of both trends in Lezghian languages). The Avar form is somewhat strange phonetically (*s:in would be regularly expected), so it may be a borrowing from some unknown early source.
Notes: Reconstructed for the PEC level. In the reduplicated form (*sūls_ūli), which is of course common EC, there occurred different assimilations (in PA: **sus:ur > *susur; on PTs see above; in Lak. *sus:- > sus; in PD *suls:ul > *s:u(l)s:ul, and likewise in PL - although we must say that in PL also *solsol or *s:ols:ol may be reconstructed; evidence for *s: is present only in the non-reduplicated stem *s:ol).
The root is also widely spread in Turkic, Mongolian, Fenno-Ugric and Kartvelian languages (see Abayev 1979, 194). Abayev (ibid.) suggests that its original source was somewhere in Central Asia. However, the time of its appearance in the Caucasus is not clear, and correspondences between languages strongly suggest its archaic character (of course if we exclude obvious late loans like the isolated Chech. sula 'oat' < Osset. syl or Georg. svila). It is interesting that only in EC languages (not in Turkic or any other mentioned above) there exists a reduplicated stem (which obviously was the source of Osset. sysyly 'a k. of weed'). Summing up, the root is obviously a "Wanderwort"; but it is evidently very ancient on Caucasian ground.
Notes: A cultural term, revealing several irregularities (-ṭ- in Chir., devoiced *-t in PL, unclear vocalism), possibly due to old interlingual borrowings. Not very reliable.
Notes: Reconstructed for the PEC level. The resonant *-j- is reconstructed to account for the development *ś- > *sṭ- in PN. The root may have also contained a laryngeal (if the PL pharyngealisation is not expressive in this case).