Comment: The verb is preserved only in Agul /Burshag dialect/; the noun 'sieve' in Agul is a later derivate (Rich. sirfel, Bursh., Burk. zerfel, Fit. zarfil). All other languages (except Tsakhur) have preserved only the PL derivate *s:iʎ:ʷ 'sieve' (besides the forms listed cf. also Lezg. Khl. sef, Akht. sef, Tab. Düb. sif:i, Khür. s:if:, Rut. Khniukh suf, Ams. süx). Most languages reflect the form with an old preverb *s:-; the only exception is Tsakh. w-ex:ʷa ( < *w-äʎ:ʷaj with another prefix and Ablaut), pointing to an original verbal root *ʔiʎ:ʷV ( / *ʔäʎ:ʷV) 'to sift'.
Comment: Cf. also Lezg. Khl. saw, Ag. Bursh. mus:u (pl. mus:uw-ur). Ag. mus:u < *mi-s:iw with a secondary *mi- (none of the other languages has any trace of it), obviously under the influence of the semantically close miḳi < *mɨƛ̣ɨj (*wɨnƛ̣ɨj) 'malt'. The obl. base can be reconstructed as *s:iwɨ-, cf. Lezg. sawu-, Rut. suji- ( < *siwɨ-).
Meaning:1 light 2 small 3 transparent, pure; liquid
Lezghian:q̇e-zil 1
Agul:ḳisal-f (?) 1
Rutul:sɨl-dɨ 1
Kryz:siläd 2
Budukh:sala-l 3
Archi:s:ala 1
Comment: Cf. also Lezg. Khl. q̇e-sɨl (Lezg. q̇e- is an archaic adjective prefix; it can not, however, correspond to Ag. ḳi- - that is why the Ag. form still requires an explanation). Shakhdagh > Khin. sulu 'liquid' (with the same specific semantic change as in Bud.).
Meaning:1 front part of leg 2 spine 3 back 4 elbow 5 edge
Lezghian:sin 5
Tabasaran:sn-ič: (Kand.)
Rutul:sɨn 1
Tsakhur:(jɨq̇Iana) sini 2
Kryz:sini 2
Archi:s:on 3
Udi:sun 4
Comment: 3d class in Tsakh., but 4th class in Rut., Arch. and Kryz. Obl. stem *s:ɨna-, cf. Lezg. sine-, Rut. sɨ́nɨ-. The Tab. word ( < *sin-ič:) contains an expressive suffix and means 'wedge' (for the meaning cf. Lak. s:an : s:ana).
Comment: Obl. base *s:ɨwɨ-, cf. Lezg. sivi-, Ag. sivi-, Tsakh. sivɨ-. Cf. also (for the reflexes of *s:) Tab. Düb. už̌u, Ag. Bursh. s:iw. Tab. and Ud. reflect a metathesised form *ʔɨs:ʷ (which may be archaic in the light of external evidence). Belongs to the 4th class in all class-distinguishing languages. Rutul has not preserved the root independently; but probably a trace of it has stayed in xe-sɨ-m ( < xex-sɨ(w)-) 'face' ('nose'+'mouth').
Comment: The Tab., Ag. (cf. also Bursh. s:ul), Tsakh., Arch. and Ud. forms go back to the simple stem PL *s:oIla (obl. base *s:oIlä- > Tab. sula-, Ag. sule-). The Lezg. and Rut. forms reflect a derivate *s:oIl-ḳ, while the Shakhdagh languages reflect *s:oIl-Vk > *s:oIkVl (probably with secondary deglottalisation of *-ḳ). Cf. similar derivates in Lak. and in Nakh. 3d class in all class-distinguishing languages.
Comment: Cf. also Lezg. Khl. sil, Ag. Bursh. s:ül. 3d class in Rut., Tsakh. and Bud., but 4th class in Kryz. As in the name of wheat (*q:ol), there is a variation of pure and pharyngealized *oI (the variant *s:oIl is reflected in Lezg. sil, Ag. s:ül, and also Kryz. and Bud.). Several forms reflect the reduplicated variant *s:ols:ol (inherited from PEC). Finally, there must have existed a suffixed (diminutive) form *s:ol-Vḳ, which was metathesized ( > *s:oḳol) and reflected in Rut. and Tsakh. (cf. in Rut. a parallel form from *s:ols:ol : Ikhr. sisel "millet").
For the simple stem *s:ol an oblique base *s:ole- can be reconstructed on basis of Lezg. sili- and Ag. (Rich.) süle-, Fit. suli-, Bursh. s:üle-, Burk. sule-.
Comment: Cf. also Lezg. Khl. sew (with metathesis of labialisation < *s:ʷeʔ), Tab. Düb. š̌eʔe, Tsakh. Gelm. soj. Strong *s:ʷ- is suggested by Ud. š- (all other languages present do not show a clear distinction of *sʷ and *s:ʷ). 3d class in all class-distinguishing languages. The oblique base may be safely reconstructed as *s:ʷeʔ-rV-, cf. Lezg. sewre-, Tab. Kand., Düb. š̌eʔra-, Rut. sirɨ-. The oblique base superseded the direct one in Kryz. and Bud.
Meaning:1 straw 2 ear and stalk of mown corn, bread
Lezghian:sam 1
Tabasaran:š̌um 1
Agul:s:um (Bursh.) 1
Rutul:sum 1
Tsakhur:som (Mishl.) 1
Kryz:sum 1
Budukh:sum 1
Udi:šIum 2
Comment: Obl. stem not clear. 4th class in all class-distinguishing languages. See Бокаpев 1961, 59; Лексика 1971, 176; Гигинейшвили 1977, 71; Талибов 1980, 271.
Comment: In Kryz. and Bud. the root has merged with PL *qIomoṭ (q.v.), yielding Kryz., Bud. šimiṭ (although the initial š- is still not clear). 4th class in all class-distinguishing languages.
Comment: Cf. also Tab. Düb. š̌um-ʒaʒa (Tab. zaz, Düb. ʒaʒa 'thorn'), Tsakh. Tsakh. suna. 3d class in Rut., but 4th class in Tsakh. -m in Tab. is secondarily labialised (š̌um < *š̌un with assimilation). There is not enough evidence for distinguishing between *s:ʷ- and *sʷ- in this root (external evidence is in favour of *s:ʷ-).
Comment: Cf. also Lezg. Nüt. gür, Akht. žur 'half'; for the obl. stem cf. also Tab. š̌uri-, Rut. sura-. Several languages reflect also the PL derivate *s:ʷɨra-n : cf. Tab. Kand. ž̌uran 'half', Düb. ž̌aran 'side', Ag. Bursh. zuran id. (cf. characteristic voicing of the original strong fricative in dialectal forms.) The Arch. form also reflects an -n-derivate, but is phonetically and semantically very archaic: it goes back to *wars:ʷɨ-n probably reflecting an original form *rɨws:ʷa- (< *rɨms:ʷa- = Lak. niws), while in all other languages the metathesis went in another direction (*rɨws:ʷa- > *s:ʷɨra-).
Comment: Both the Lezg. and the Ag. forms are pluralia tantum. In Lezg. cf. also Akht. žur, Nüt. gürer id. A loan from some Lezg. dialect may be Ud. (Vart.) zurul 'wild pear' (although there are some doubts as to the meaning of the word).
