There are two documented expressions for 'all':

    1) li-der-(u) [class 1, 2, 4] / lo-der-(u) [3, 5], which is quoted as the only equivalent for 'all (omnis)' in [Isakov & Khalilov 2012: 161; Bokarev 1961: 173]. The plural forms 'all (omnis)' are given as regular li-der-ar [human pl.] / lo-der-ar [non-human pl.] in [Isakov & Khalilov 2012], but as li-der-ol in [Bokarev 1961]. This is the present participle in -der of the generic verb 'to be' [van den Berg 1995: 99 ff.], i.e., 'all' as 'whoever/whatever) being'. It should be noted that in [van den Berg 1995: 314], li-der-u is translated as 'every, each'.

    2) seh / setʼ (with unclear distribution of variants), which is quoted for 'all (omnis / totus)' in [Isakov & Khalilov 2001: 148; van den Berg 1995: 332]. The examples are: “She made them take off all their upper clothes” [van den Berg 1995: 246], “All the animals have come” [Isakov & Khalilov 2001: 148].

    We have to treat li-der- and seh as synonyms.

    Several terms for 'all' are listed in [Khalilov 1995: 300, 401]. Out of them, the adjective qʼacʼː-o is the most frequently used one, as follows from browsing through [Khalilov 1995]. Cf. the examples for attributive use: “to find out all the circumstances” [Khalilov 1995: 31], “In summer, the cattle are all in the mountains”, “All the books are mine” [Khalilov 1995: 153]. Non-attributive use: “There was room for all (of them)” [Khalilov 1995: 118], “All (i.e., everybody) have got their shares” [Khalilov 1995: 215], “All (i.e., everything) will be all right” [Khalilov 1995: 217].

    A second candidate is the non-inflected form setʼ {сетI} [Khalilov 1995: 228] with the following examples quoted in the main entry: “All the children have come”, “All (i.e., everybody) are singing the song”, “All (i.e., everybody) have stopped talking”.

    A third candidate is the adjective gäːh-iy-o {га̄ьгьийо} [Khalilov 1995: 59, 401; Madieva 1965: 103], but normally it is used in non-attributive function: “All (of them) have gone to the cinema”, “All (of them) have books” [Khalilov 1995: 59]. gäːh-iy-o is derived from gäh-iy-o 'existing' [Khalilov 1995: 58] with the iterative infix -a- (gäh-iy-o is the participle from the auxiliary present stem gey 'to be' [Khalilov 1995: 59], although the vowel change e ~ ä is abnormal).

    Distinct from siyo-nazu 'everything, each' [Khalilov 1995: 229], sukʼo-nazu 'everybody, each' [Khalilov 1995: 233] with interrogative siyo 'what', sukʼo 'who' q.v. and the special generalizing element -na-zu.

    Other candidates are seh 'all' [Khalilov 1995: 228] and qʼacʼː-o 'all' [M. Khalilov, p.c.], but these forms are apparently more marginal than gälːö, because seh and qʼacʼː-o are not mentioned in [Kibrik & Testelets 2004].

    Distinct from sukʼo-nazu 'each, every' [Kibrik & Testelets 2004: 249] with interrogative sukʼo 'who' q.v. and the special generalizing element -na-zu.

    A second, apparently more marginal candidate is naːsi-n {на̄син} 'all (omnis) / all (totus)' [Khalilov 1999: 197, 312, 440; Imnaishvili 1963: 130], derived from the interrogative pronoun naːsi 'which (one)?' [Khalilov 1999: 197] with the particle -n (for which cf. [Imnaishvili 1963: 265]).

    Borrowed pronouns are used as well: kinaw-ni 'all' [Khalilov 1999: 140, 312; Imnaishvili 1963: 133] < Avar kina-w-go 'all'; šinaw 'every, all' [Khalilov 1999: 291; Imnaishvili 1963: 133] < Avar šːina-w 'every, all'.

    Kwantlada Khwarshi: golːu-č {голлуч} 'all' [Karimova 2014; Khalilova 2009: 165]. Consists of the present participle golːu (< *gol-y-u) from the auxiliary present stem gol-i 'to be' [Khalilova 2009: 182] plus the polyfunctional suffix -č, which, in particular, forms collective numerals [Khalilova 2009: 177].

    Replacements: {'to be' > 'all'} (Bezhta proper, Tlyadal Bezhta, Khwarshi).

    Reconstruction shape: Correspondences seem regular except for some vowel fluctuation.

    Semantics and structure: Verbal participle.

    Distinct from čʼečʼa {чIечIа} 'soot; thin crust of ice over snow' [Khalilov & Isakov 2005: 396].

    Quite differently in [Kibrik & Kodzasov 1990: 207], where čʼečʼa is proposed as the basic word for 'ashes', whereas noƛu (sic!) is specified as 'fine ashes (пепел)'. Apparently a mass of errors.

    Kwantlada Khwarshi: yõƛu {йо^нлIу} 'ashes' [Karimova 2014].

    Reconstruction shape: Correspondences seem regular, except for the occasional shift of nasalization between the vowel and the initial glide.

    Semantics and structure: Primary substantive root.

    Distinct from qal {хъал} 'peel, bark' [Isakov & Khalilov 2001: 166, 207; Bokarev 1961: 167], borrowed from Avar qːal 'peel, bark'.

    Distinct from qʼeqʼel-ba {къекъелба} 'birch bark' [Isakov & Khalilov 2001: 98, 190], final -ba is the plural exponent; it must be noted that in [Isakov & Khalilov 2001: 101] the dialectal variant ƛʼeqʼel-ba {кьекъелба} (village Garbutl) is quoted, which seems to be a graphical corruption (Cyrillic {ь} for {ъ}).

    There are three documented terms for 'bark':

    1) derived term y=ic-al-o;

    2) primary stem beš {беш} with polysemy: 'bark / shell / skin' [Khalilov 1995: 45, 311; Madieva 1965: 150];

    3) borrowed term qal {хъал} with polysemy: 'bark / peel / skin' [Khalilov 1995: 259, 311; Madieva 1965: 190] < Avar qːal 'peel, bark'.

    Provisionally we choose y=ic-al-o as the basic Bezhta proper term, because there are two textual examples for it: “dry bark”, “to peel bark from tree” [Khalilov 1995: 123]. Only one example is available for qal: “thin bark” [Khalilov 1995: 259], and no examples for beš. It should be noted that beš is the basic term for 'human skin' q.v.

    Distinct from ƛʼeqʼa {кьекъа} 'patch (a piece of cloth)' [Khalilov 1995: 163].

    In [Comrie & Khalilov 2010: 412], ricalo is also quoted as one of the Khoshar-Khota terms for 'bark' (directly corresponds to Bezhta proper y=ic-al-o).

    There are two Tlyadal terms for 'bark', quoted in [Kibrik & Kodzasov 1990] as synonyms: ƛʼeqʼä and qar. Additionally, M. Khalilov (p.c.) quotes the Avar loanword qal 'bark'.

    A second term is the borrowing qal {хъал} with polysemy: 'bark / peel / layer, coat / skin' [Khalilov & Isakov 2005: 358, 456] < Avar qːal 'peel, bark'.

    Distinct from ƛʼiqʼin {кьикъин} 'birch bark' [Khalilov & Isakov 2005: 228; Kibrik & Kodzasov 1990: 97].

    Distinct from the borrowed term qal {хъал} with polysemy: 'bark / peel / layer, coat / skin' [Khalilov 1999: 251, 334] < Avar qːal 'peel, bark'.

    Distinct from the more specific term meču {мечу} 'birch bark' [Khalilov 1999: 189, 334].

    Distinct from the more specific term čʼita 'birch bark' [Kibrik & Kodzasov 1990: 97].

    Kwantlada Khwarshi: qʼʷelʸ {къвел} 'bark' [Karimova 2014; Khalilova 2009: 16, 68].

    'Bark' is a less stable item in East Tsezic. The match between Hunzib r=ič-ul and Bezhta proper y=ic-al-o plus Khoshar-Khota Bezhta r=ic-al-o, derived from the verb 'to peel' (< Proto-Tsezic *class=ič- 'to peel, take off skin' [NCED: 265]) with the non-productive l-suffix, suggests that this deverbative should be the Proto-East stem for 'bark'. In the Khoshar-Khota-Tlyadal cluster, it was superseded with *ƛʼeqʼV, whose original Proto-Bezhta meaning was 'patch' (as follows from the Bezhta proper data). Further to the suffixed Hinukh stem ƛʼiqʼi-n 'birch bark'. No East Caucasian etymology. In many lects, inherited forms tend to be superseded with the Avar loanword (Hunzib proper, Bezhta proper, Hinukh, Kidero Dido, Sagada Dido, Inkhokwari Khwarshi).

    Replacements: {'bark' > 'crust'} (Hunzib proper), {'to peel' > 'bark'} (Hunzib proper, Bezhta proper, Khoshar-Khota Bezhta), {'skin' > 'bark'} (Bezhta proper), {'patch' > 'bark'} (Khoshar-Khota Bezhta, Tlyadal Bezhta).

    Reconstruction shape: Correspondences seem quite regular.

    Semantics and structure: Primary substantive root.

    Distinct from ɬili ~ ɬüli {лъили, лъили} with polysemy: 'belly / stomach / bosom (as in “in one's bosom”)' [Khalilov & Isakov 2005: 248, 440], which is applicable specifically to humans and seems to be more rare than aχ.

    Distinct from ɬara {лъара} with polysemy: 'belly / bosom (as in "in one's bosom")' [Khalilov 1999: 171, 324], which is applicable specifically to humans and seems to be used more rarely than aχ.

    Distinct from bačʼʷa-hala 'stomach' [Kibrik & Kodzasov 1990: 36].

    Kwantlada Khwarshi: ãχ ~ ãχˤ {а^нх} 'belly' [Karimova 2014; Khalilova 2009: 16]. Polysemy: 'belly / stomach'. The variant ãχˤ is from [Khalilova 2009].

    The root *ʔǝ̃χː (~ ʁ), attested as the Common Bezhta term for 'belly', seems isolated in Tsezic; for possible external comparanda see [NCED: 676].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root with polysemy: 'belly / stomach'.

    Distinct from the more specific term iχ-y-u {ихйу} 'large, massive' [Khalilov 1999: 130].

    Distinct from the more specific term iχ-a-w {ихав} 'large, massive' [Abdulaev 2014].

    Kwantlada Khwarshi: class=uqʼːˤ-u ~ class=uqʼˤ-u {лʼукъкъуI} 'big' [Karimova 2014; Khalilova 2009: 16].

    In Hinukh and Dido, *class=uqʼˤV was superseded with *class=iˤžV ~ -žː- B [NCED: 653], whose original meaning was something like 'many, numerous' vel sim., cf. sub 'many'. Apparently, parallel contact-driven introductions in Hinukh and Dido may be suspected.

    Replacements: {'big' > 'many'} (Kidero Dido, Sagada Dido).

    Reconstruction shape: Correspondences seem regular except for the root vowel in Hunzib.

    Semantics and structure: Primary adjective root.

    Kwantlada Khwarshi: kʼuca ~ kʼɨca {кIуца} 'bird' [Karimova 2014; Khalilova 2009: 7]. In [Karimova 2014], only kʼuca is quoted.

    The formal match between Bezhta mihna 'bird' and Hinukh mihna 'small bird / young of animal' makes *mihnV the main candidate for the status of Proto-Tsezic. In Hunzib, this stem is reflected as mina 'young of animal (incl. nestling)' [Isakov & Khalilov 2001: 123]. Probably one should reconstruct *mihnV with Proto-Tsezic polysemy: 'small bird / young of animal', retained in Hinukh, but eliminated in different ways in Bezhta and Hunzib. No North Caucasian external etymology for Proto-Tsezic *mihnV is available. It must be noted that theoretically *mihnV with the meaning '(small) bird' can only be an East Tsezic feature, whereas Hinukh mihna 'small bird / young of animal' arose under the influence of East Tsezic.

    In other languages, three different forms for '(small) bird' are used; each of them seems to be isolated in Tsezic:

    1) Hunzib čʼeq < Proto-Tsezic *čʼeq < North Caucasian 'a kind of small bird' [NCED: 1105].

    2) Dido aʁi with a very weak external (Avar) comparandum [NCED: 511]; pace [NCED: 511], Hunzib ãχ 'bird' does not exist.

    3) Khwarshi kʼeca < Proto-Tsezic *kʼɨca (~ -i-) < North Caucasian 'a kind of small bird' [NCED: 442].

    Replacements: {'a kind of small bird' > 'bird'} (Hunzib, Khwarshi).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root with polysemy: 'small bird / young of animal'.

    Distinct from gɑ̃žu y=ɑhu 'to bite', applied specifically to dogs [van den Berg 1995: 297], literally 'to take away the fang' with gɑ̃žu 'animal fang' and class=ɑhu 'to take away, take off, seize' [van den Berg 1995: 284].

    Cf. the parallel expression sila y=aʁo {сила йагъал} 'to bite off' [Khalilov 1995: 230], literally 'to take out the tooth' with class=aʁo 'to take out' [Khalilov 1995: 105].

    Kwantlada Khwarshi: han {гьа^нна} 'to bite' [Karimova 2014; Khalilova 2009: 255]. Applied to both humans and animals (e.g., a dog).

    Replacements: {'to beat the tooth to/for smth.' > 'to bite'} (Hunzib); {'to put the tooth' > 'to bite'} (Bezhta).

    Reconstruction shape: The irregular correspondence -n (Dido, Khwarshi) ~ -l (Hinukh) remains inexplicable, although there is little doubt that the attested West Tsezic forms are related to each other. If the external etymology proposed in [NCED: 625] is correct (< North Caucasian *=iʡʷVl), *heˤl- is the original Proto-Tsezic variant retained in Hinukh, although it is actually *heˤn- that shows a broader distribution (Dido, Khwarshi).

    Semantics and structure: Primary verbal stem.

    Kwantlada Khwarshi: kaˤba {каIба} 'black' [Karimova 2014; Khalilova 2009: 8].

    1) *cʼǝdV-l- [NCED: 379], meaning 'black' in East Tsezic (Hunzib, Bezhta), lost in West Tsezic;

    2) *kaˤba, meaning 'black' in West Tsezic (Hinukh, Dido, Khwarshi), lost in East Tsezic.

    Since *cʼǝdV-l- has promising external comparanda (Andian 'blackberry', Dargi 'black'), whereas *kaˤba has no North Caucasian etymology, *cʼǝdV-l- can be postulated with more probability as the Proto-Tsezic term for 'black'.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Adjective stem with the common l-suffix.

    Kwantlada Khwarshi: ẽqʼo ~ ẽqʼˤo {э^нкъо} 'blood' [Karimova 2014; Khalilova 2009: 5, 16]. The variant ẽqʼˤo is from [Khalilova 2009].

    Reconstruction shape: Correspondences seem regular except for the unclear element -qʼʷa ~ -qʼo in Khwarshi, which is either a unique suffix or an unknown root, compounded with *ħɔ̃y.

    Semantics and structure: Primary substantive root. The oblique stem is *ħẽyɑ-.

    A second candidate is qʼʷaqʼu {къвакъу} 'bone' [Khalilov & Isakov 2005: 213, 457]; the difference between ƛužey and qʼʷaqʼu is unclear, but the latter seems more marginal, because it is missing from [Kibrik & Kodzasov 1990]. The same word with loss of labialization is documented as qʼaqʼu {къакъу}, which is glossed as 'collarbone' in [Khalilov & Isakov 2005: 211] and as 'round end of bone' in [Kibrik & Kodzasov 1990: 38].

    Distinct from ƛuza, obl. ƛuza- {лIуза}, glossed in [Khalilov 1999: 180] as generic 'bone', but marked as rarely used; on the contrary, in [Kibrik & Kodzasov 1990: 38], ƛuza is quoted as a generic term for 'tubular bone'.

    It is likely that the analysis in [Khalilov 1999] is more correct, implying that in Proto-Dido, ƛuza was the basic term for 'bone', but currently it is being superseded by qʼ(ʷ)ˤaqʼu.

    Kwantlada Khwarshi: ƛozol {лIозол} 'bone' [Karimova 2014; Khalilova 2009: 5].

    Reconstruction shape: Correspondences seem regular, except for the consonant metathesis in the oblique stem.

    Semantics and structure: Primary substantive root. The oblique stem is *ƛʷɨrV-zV-, which was eliminated, i.e., levelled after the direct stem in East Tsezic (Hunzib, Bezhta). In West Tsezic (Hinukh, Dido, Khwarshi), oblique *ƛʷɨrV-zV- was metathesized > *ƛʷɨzV-rV-, whereupon the direct stems in individual lects were levelled: Hinukh ƛužey, Kidero Dido ƛuza are back-formations; Khwarshi proper ƛazal, Inkhokwari Khwarshi ƛozol continue oblique *ƛʷɨzV-rV-. It is interesting that due to the Sagada Dido direct stem ƛurza, which directly goes back to oblique *ƛʷɨrV-zV-, this metathesis in the oblique stem *ƛʷɨrV-zV- > *ƛʷɨzV-rV- cannot be projected onto the Proto-West Tsezic level - more likely, we are dealing with late contact-driven rebuildings of the paradigm in individual West Tsezic lects. Cf. similar rebuildings of the original paradigm in the words for 'eye', 'hand' q.v.

    Distinct from the nursery word nene {нене} with polysemy: 'female breast / udder' [Isakov & Khalilov 2001: 132, 197; Kibrik & Kodzasov 1990: 22; van den Berg 1995: 321; Bokarev 1961: 162].

    Distinct from χɨmǝr {хымǝр} 'brisket' [Isakov & Khalilov 2001: 165; van den Berg 1995: 341].

    Distinct from the more rare term χomaː {хома̄}, glossed as 'breast, thoracic cage' in [Khalilov 1995: 256], not quoted in [Madieva 1965].

    Distinct from tʼum 'female breast' [Khalilov & Isakov 2005: 340, 432; Kibrik & Kodzasov 1990: 22].

    Distinct from the nursery word kiki {кики} 'female breast' [Khalilov 1999: 139, 317; Kibrik & Kodzasov 1990: 22].

    Distinct from nursery kaka 'female breast' [Bokarev 1959: 147].

    Distinct from the nursery word koko 'female breast / nipple' [Kibrik & Kodzasov 1990: 22; Bokarev 1959: 147, 149, 151].

    Kwantlada Khwarshi: humo-lokʼo {гьу^нмолокIо} 'breast' [Karimova 2014].

    Additionally, in Hinukh, the form ʁomo 'udder' is attested [Khalilov & Isakov 2005: 131]; it is probably related to the discussed forms and therefore could represent borrowing in the opposite direction: Bezhta > Hinukh (although no such forms are attested in Bezhta).

    A second candidate for Proto-Tsezic 'breast' is *χeru A [NCED: 465], which means 'breast' in all East Tsezic lects (can be safely postulated as the Proto-East Tsezic term for this meaning) and in one of the West Tsezic lects, namely Khwarshi proper, where the compound ħele-lokʼʷa, literally '*χeru-heart' is used. At first sight, *χeru has the advantage over *χɨmV(-rV) from the distributional point of view, but in fact it is hard to suppose that *χeru was the Proto-West Tsezic term for 'breast', which only survived in Khwarshi proper, having been superseded in the rest of the lects.

    Since *χɨmV(-rV) possesses very good external North Caucasian comparanda with the meaning 'breast' [NCED: 829], we fill the Proto-Tsezic slot with *χɨmV(-rV). The Proto-Tsezic or at least Proto-West Tsezic meaning of the competing term *χeru is unclear. Its 'non-breast' semantics is attested in Dido proper (ħiro 'shoulder'); on the other hand, external North Caucasian comparanda point to a meaning like 'udder' [NCED: 465]. The use of *χeru in the Khwarshi proper compound for 'breast' remains inexplicable, but we suppose that it is a secondary phenomenon.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root, optionally modified with the suffix -rV.

    Cf. the more specific verb ɬehe {лъегьа} 'to burn (intrans.)' [Isakov & Khalilov 2001: 197; van den Berg 1995: 315; Bokarev 1961: 160, 173], application is unknown.

    Kwantlada Khwarshi: class=okʼ ~ class=okʼ-χ {локIа, локIха} 'to burn' [Karimova 2014; Khalilova 2009: 85, 246, 275, 338].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Either the labile verbs *class=ekʼʷV, with polysemy 'to burn (intrans.) / to burn (trans.)' already in Proto-Tsezic, or the causative *class=ekʼʷV-l is to be reconstructed for Proto-Tsezic (later, the Proto-Tsezic causative exponent -l was replaced by synchronic causative suffixes in Hunzib and Khwarshi).

    Kwantlada Khwarshi: munu {муну} 'fingernail' [Karimova 2014].

    Reconstruction shape: Correspondences seem regular except for the nasal assimilation l > n in Inkhokwari Khwarshi and the vowel i in Bezhta due to influence of the oblique stem.

    Semantics and structure: Primary substantive root. The oblique stem is *mɨˤʫa A.

    Distinct from the specific term žanʁi {жангъи} 'fog' [Khalilov & Isakov 2005: 166, 535].

    Cf. also kutʼi, which is glossed with polysemy 'fog / smoke' in [Kibrik & Kodzasov 1990: 205], but only as 'smoke' in [Khalilov & Isakov 2005: 208].

    Cf. also gutʼ, which is glossed with polysemy 'fog / smoke' in [Kibrik & Kodzasov 1990: 205], but only as 'smoke' in [Khalilov 1999: 87].

    Kwantlada Khwarshi: as {ас} with polysemy: 'cloud / sky' [Karimova 2014; Khalilova 2009: 110, 265].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Kwantlada Khwarshi: class=ucʼː-u {луцIцIу} 'cold' [Karimova 2014; Khalilova 2009: 104, 301]. Applied to both objects and weather.

    Reconstruction shape: Some phonetic outcomes are irregular: čʼ instead of expected cʼ in Hinukh, cʼ instead of expected čʼ in Khwarshi.

    Semantics and structure: Participle with the regular y-suffix from the stative verb *class=ɔčʼ(V)- 'to be cold, become cold', which is retained as a synchronic verb at least in Kwantlada Khwarshi.

    Distinct from n=ǝː [class 1] / n=iː [2] / n=uː [4, pl.] {нǝа, ниа, нуа} 'to come (here)' [Isakov & Khalilov 2001: 226; van den Berg 1995: 33, 78, 322, 353; Bokarev 1961: 162, 179]. Originates from something like *nV=class=ǝ; initial n= is a fossilized directional prefix [van den Berg 1995: 353].

    Semantic or pragmatic difference between class=ãqʼe and n=ǝː is unclear, the latter seems to be missing from the main section of [Isakov & Khalilov 2001].

    Cf. with another directional prefix: g=ǝː [1] / g=iː [2] / g=uː [4, pl.] {гува} 'to come down (of precipitation)' [Isakov & Khalilov 2001: 56; van den Berg 1995: 299, 353]; < *gV=class=ǝ; g= is a fossilized directional prefix 'down' [van den Berg 1995: 353]. It should be noted that for the Naxada dialect, g=ǝː is glossed as generic 'to come (here)' [van den Berg 1995: 299, 353].

    Distinct from the imperative t=ǝs [1] / t=is [2] / t=us [4, pl.] 'come with me! / come to me!' [van den Berg 1995: 336, 353; Bokarev 1961: 165] with another fossilized directional prefix, t=.

    There are two verbs for 'to go' quoted in [Khalilov 1995: 329]: =õqʼo and g=Vh. We treat them as synonyms.

    There are two verbs for 'to go' quoted in [Kibrik & Kodzasov 1988]: =oqʼo and g=Vː. We treat them as synonyms.

    There are two verbs for 'to come' quoted in [Khalilov & Isakov 2005: 503]: class=aqʼe and noχ. The exact difference between them is unknown; browsing through available sources confirms that both are used quite frequently. We are obliged to threat them as synonyms.

    A second candidate, found in [Khalilov 1999], is neχ 'to come'. We treat both verbs as synonyms.

    The difference between two documented verbs for 'to come', =atʼiqʼ & =iχʷ, is unclear; we have to treat them as synonyms.

    Kwantlada Khwarshi: class=otʼqʼ {отIкъа} 'to come' [Karimova 2014; Khalilova 2009: 181]. As in other Khwarshi varieties, there are two known verbs for 'to come': class=otʼqʼ and class=uχ ~ class=uχˤ {уха} [Karimova 2014; Khalilova 2009: 32, 43]. Browsing through textual examples in [Khalilova 2009] suggests that (1) =otʼqʼ and =uχ are complete or almost complete synonyms in the meaning 'to come'; (2) =otʼqʼ is used much more frequently than =uχ.

    Cf. an example in which two verbs are used in parallel constructions: "...the Bagwalals came (=otʼqʼ), the Echedas came (=otʼqʼ), other people also came (=uχ) from around from other villages ..." [Khalilova 2009: 95].

    Other attested examples for =otʼqʼ 'to come' are: e.g., "That man came to our village" [Khalilova 2009: 42]; "My children came" [Khalilova 2009: 44]; "The manual machine gunner who was in Manchuria came back having hurt his legs" [Khalilova 2009: 75]; "The giant came there while they were sitting under the tree" [Khalilova 2009: 77]; "The wolf came near the apple tree" [Khalilova 2009: 87]; "When he came near the house, and before going inside, he ..." [Khalilova 2009: 90]; "The eldest (girl) came" [Khalilova 2009: 99]; "There, the man who sells watermelon came" [Khalilova 2009: 116].

    On the contrary, available examples for =uχ 'to come' are rather scant; cf. "Hey people, come, there is something in my eye, take it out" [Khalilova 2009: 73], "Put the food over there, I will come to eat" [Khalilova 2009: 116], "Then the wolf went from there to the donkey" [Khalilova 2009: 118], "When (donkey) went from uphill down the hill, (donkey) met a horse" [Khalilova 2009: 120].

    Apparently, already in Proto-East Tsezic *class=ãqʼV started to compete with the non-standard verb *dir=class=VH [NCED: 1016], where *dir= is a directional prefix. In modern East Tsezic lects both verbs for 'to come' function as synonyms with unclear distribution, although *class=ãqʼV- seems to be more common. In West Tsezic, *=VH is not attested; the original Proto-Tsezic meaning of *=VH is unclear.

    A similar situation is observed in the West Tsezic group. Apparently, already in Proto-West Tsezic *class=ãqʼV- began to compete with the verb *class=uχːʷ- B [NCED: 666], which is sometimes attested with the fossilized directional n-prefix (Hinukh, Dido). In modern West Tsezic lects, both verbs for 'to come' function as synonyms with unclear distribution, although *class=ãqʼV- seems to be more common. In East Tsezic, *=uχːʷ- is not attested; the original Proto-Tsezic meaning of *=uχːʷ- is unclear.

    Reconstruction shape: Correspondences seem regular, except for some vocalic peculiarities.

    Semantics and structure: Primary verbal root.

    Distinct from euphemistic class=ikʼe-ɬ {бикIелъа} with polysemy: 'to get lost, disappear / to die' [Khalilov & Isakov 2005: 84].

    Distinct from rarer, stylistically marked synonyms: rude geg {гега} with polysemy: 'to get squashed / to get broken / to die' [Khalilov 1999: 81], rude ɬiqu {лъихъа} with polysemy: 'to dry (intrans.) / to die' [Khalilov 1999: 175], rude ʁutʼ {гъутIа} 'to die (only of animals?)' [Khalilov 1999: 94], polite kec {кеца} with polysemy: 'to lie / to sleep / to die' [Khalilov 1999: 138].

    Kwantlada Khwarshi: class=uh {угьа} 'to die' [Karimova 2014; Khalilova 2009: 6, 26].

    Replacements: {'to die' > 'to get spoilt'} (Hunzib proper, Bezhta proper, Tlyadal Bezhta, Hinukh, Kidero Dido); {'to die' > 'to get finished'} (Bezhta proper).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary verb; polysemy: 'to die / to get spoilt' can be reconstructed for Proto-Tsezic.

    Distinct from the loanword baħli {бахIли} 'hunting dog' [Isakov & Khalilov 2001: 33; van den Berg 1995: 286; Bokarev 1961: 150], borrowed from Avar baħri 'stray dog'.

    Distinct from the loanword bähri {баьгьри} 'stray dog' [Khalilov 1995: 42], borrowed from Avar baħri 'stray dog'.

    Distinct from the loanword baħri {бахIри} 'hunting dog / stray dog / vagrant' [Khalilov 1999: 40], borrowed from Avar baħri 'stray dog'.

    Kwantlada Khwarshi: ʁʷˤe {гъвеI} 'dog' [Karimova 2014; Khalilova 2009: 16].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Distinct from the compound sõƛe {со^нлIа} 'to sip' [Isakov & Khalilov 2001: 149].

    A synonym is found in the compound verb hi-ƛ {гьилIа} 'to drink / to sip' [Khalilov & Isakov 2005: 143, 490], marked by Khalilov & Isakov as a rarely used form. Historically, a compound: hi-ƛ, where the second element ƛ is the verbum dicendi *ʔiƛV- (> Hinukh eƛi 'to call, cry / to say' [Khalilov & Isakov 2005: 405]). For Hinukh complex verbs in -ƛe denoting sounds, see [Forker 2013: 322]. Note the morphological difficulty with the final vowel: the independent verb is eƛi, the second element of sound denoting verbs is -ƛe, the second element of the aforementioned verb 'to drink' is -ƛ.

    A second candidate is cʼod {цIода} 'to drink' [Karimova 2014] without examples.

    A second Inkhokwari candidate is cʼod {цIода} 'to drink' [Karimova 2014; Imnaishvili 1963: 187, 240]. No textual instances were found either for χu-ƛ or for cʼod, but, through analogy with other Khwarshi dialects, we suppose that χu-ƛ is the basic verb.

    Kwantlada Khwarshi: χu-ƛ {хулIа} 'to drink' [Karimova 2014]. The stem is χu-ƛ, not χu-ƛV, cf. the diagnostic present form χu-ƛ-še [Khalilova 2009: 184]. Browsing through [Khalilova 2009] proves that χu-ƛ is the most generic and frequently used verb for 'to drink'. Cf. the attested examples: "The mother is drinking tea now" [Khalilova 2009: 184], "The father usually drinks tea in the morning" [Khalilova 2009: 187], "Children usually drink milk" [Khalilova 2009: 189], "Don't let him drink my juice" [Khalilova 2009: 252], "Can I drink the milk?" [Khalilova 2009: 361], "Father bought beer to drink" [Khalilova 2009: 418].

    A second Kwantlada candidate is cʼod {цIода} [Karimova 2014], but examples from [Khalilova 2009] suggest that its real meanings are more specific, e. g. 'to drink alcohol' etc. Cf. the examples: "When everybody smoked (cʼod) the cigarettes, this devil began to choke" [Khalilova 2009: 165], "Our neighbor had been drinking year after year" [Khalilova 2009: 201], "The doctor made the mother make the boy drink the medicine" [Khalilova 2009: 347], "the horses went to the river to drink water" [Khalilova 2009: 401], "Once they were drinking and eating" [Khalilova 2009: 449], "There was nothing. What, did we go there for drinking?" [Khalilova 2009: 457].

    Reconstruction shape: Correspondences seem regular except for the Dido form with irregualr ħ- (which is a normal outcome of *χ, not *χː) and -a-. Despite this, the Dido verb can hardly be separated from other Tsezic forms.

    Semantics and structure: Originally a compound of *χːʷi- '?' and the verbum dicendi *ʔiƛV (cf. sub 'to say'). Strictly speaking, the original meaning of *χːʷi- remains unclear. Outside of this compound, the suffixed root *χːʷi- is attested at least in Hunzib with the meaning 'to drink (alcohol), get drunk' that should point to Proto-Tsezic meaning 'to drink'. On the other hand, the Tsezic compound pattern of -ƛV is characteristic just for sound denoting ("ideophonic") verbs; *χːʷi-ƛV 'to drink' is the main exception from this semantic group.

    A second, apparently less common candidate is ɬiq-aː-si {лъихъа̄си} with polysemy: 'dry / dead (of animal)' [Khalilov 1999: 175], past participle from ɬiqu {лъихъа} 'to dry (intrans.) / to die (rude)' [Khalilov 1999: 175].

    Kwantlada Khwarshi: quqː-u {хъухъхъу} 'dry' [Karimova 2014].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: *qoqV is the verb 'to be(come) dry', the adjectival meaning 'dry' in modern lects is expressed by various synchronic participles.

    Kwantlada Khwarshi: ãha ~ ãˤha {а^нгьа} 'ear' [Karimova 2014; Khalilova 2009: 16, 17]. The pharyngealized variant ãˤha is from [Khalilova 2009].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root; the West Tsezic forms go back to *ʔãχi-ya with a fossilized plural suffix. External North Caucasian comparison suggests that the Proto-Tsezic root is to be analyzed as *ʔã-χa, where final -χa is apparently a fossilized plural exponent.

    Distinct from the more marginal čʼabar {чIабар} with polysemy: 'soil / clay / earth floor' [Isakov & Khalilov 2001: 178, 202], borrowed from Avar čʼabˈar 'earth floor'.

    Distinct from dunyal ~ duniyal {дунйал} 'world, universe' [Isakov & Khalilov 2001: 76; van den Berg 1995: 295], borrowed from Avar dunˈiyal 'world / life / weather', ultimately from Arabic dunyaː 'earth, world'.

    Distinct from mäče {маьче} 'earth, land / plot of land' [Khalilov 1995: 193; Madieva 1965: 176].

    Distinct from two loanwords: dunːal ~ duniyal {дуннал} 'world, universe / weather' [Khalilov 1995: 93] (< Avar dunˈiyal 'world / life / weather', ultimately from Arabic dunyaː 'earth, world'); raƛʼ {ракь} 'earth, dry land / country' [Khalilov 1995: 222; Madieva 1965: 183] (< Avar raƛʼː 'earth, soil, land').

    Distinct from moči {мочи} 'earth, plot of land' [Khalilov & Isakov 2005: 269].

    Distinct from two loanwords: duniyal ~ dunyal {дунийал, дунйал} 'world, universe / weather' [Khalilov & Isakov 2005: 165] (< Avar dunˈiyal 'world / life / weather', ultimately from Arabic dunyaː 'earth, world'); raƛʼ {ракь} 'earth, dry land / country' [Khalilov & Isakov 2005: 299] (< Avar raƛʼː 'earth, soil, land').

    Distinct from moči {мочи} 'earth, plot of land / place' [Khalilov 1999: 192].

    Distinct from two loanwords: dunyal {дунйал} 'world, universe / weather' [Khalilov 1999: 116] (< Avar dunˈiyal 'world / life / weather', ultimately from Arabic dunyaː 'earth, world'); raƛʼ {ракь} 'earth, dry land / country' [Khalilov 1999: 213] (< Avar raƛʼː 'earth, soil, land').

    Kwantlada Khwarshi: čʼido {чIидо} with polysemy: 'soil / ground / territory' [Karimova 2014; Khalilova 2009: 93, 396]. Cf. the example: "When they could not bear the hunger, they began to eat the earth" [Khalilova 2009: 396].

    1) East Tsezic *mɨsV [NCED: 1055] with polysemy: 'soil / clay'; lost in West Tsezic. Its North Caicasian comparanda mean 'big stone, rock' (Andian), 'mountain pasture' (Lak), 'mountain, hill' (Dargi), 'mountain pasture; cattle-shed' (Lezgian).

    2) West Tsezic *čʼüdV [NCED: 385] with polysemy: 'soil / clay'; in East Tsezic this root is attested with the meaning 'soot' (Bezhta proper cʼide [Khalilov 1995: 271]). Its North Caucasian comparanda mean 'clay' (Andian), 'dirt, mud' (Dargi).

    There is neither internal Tsezic nor external North Caucasian evidence for making a choice between the two. Provisionally we fill the Proto-Tsezic slot 'earth' with *čʼüdV; in any case, neither choice would be relevant for further lexicostatistical analysis.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    There are two verbs for 'to eat' in Hunzib: transitive =ũq and objectless intransitive =eš. We treat them as synonyms.

    There are two verbs for 'to eat' in Hinukh: transitive =acʼ and objectless intransitive =iši. We treat them as synonyms.

    There are two verbs for 'to eat' in Kidero Dido: transitive =acʼ and objectless intransitive =iš. We treat them as synonyms.

    Kwantlada Khwarshi: class=acʼ {лацIа} 'to eat (trans.)' [Karimova 2014; Khalilova 2009: 21, 26]. Cf. some examples: "The she-goat ate the grass" [Khalilova 2009: 69], "The robber had nothing to eat" [Khalilova 2009: 160], "Each ate one apple" [Khalilova 2009: 178], "I don't eat such dirty leaves" [Khalilova 2009: 202].

    kok {кока} 'to eat (intrans.)' [Karimova 2014; Khalilova 2009: 6, 182]. Cf. some examples: "Put the food over there, I will come to eat" [Khalilova 2009: 116], "One time the father of these girls was eating at one man's (place)" [Khalilova 2009: 146], "Aminat's son does not eat at all" [Khalilova 2009: 201].

    NCED: 559. Distribution: All Tsezic lects, except for Sagada Dido, lexically discriminate between 'to eat (trans.)' and 'to eat (intrans.)'. The same opposition is to be reconstructed for Proto-Tsezic. The roots attested in the meaning 'to eat' in modern Tsezic lects can be summarized in the following table:

'TO EAT'	Hunzib	Bezhta	Hinukh	Kidero Dido	Sagada Dido	Khwarshi
*class=ũq ~ *class=ĩqʷ [NCED: 559]	trans.	trans.
*class=acʼ [NCED: 1017]			trans.	trans.	(in)trans.	trans.
*class=eš A	intrans.	intrans.	intrans.	intrans.
*kUkV A [NCED: 207]						intrans.

    The verb *class=eš A (absent in [NCED; TsezEDb]) can be safely reconstructed as the Proto-Tsezic equivalent for 'to eat (intrans.)'. It was superseded by *=acʼ 'to eat (trans.)' in Sagada Dido and by *kǝkV '?' in Khwarshi.

    The original meaning of *kUkV A, only attested in Khwarshi as 'to eat (intrans.)', is unclear. Pace [NCED: 207], it must be separated from Hinukh kiki {кика} 'to feed / to rear, raise', kikzi {кикзи} 'rearing' [Khalilov & Isakov 2005: 202; Forker 2013: 334], Kidero Dido kiki {кика} 'to rear, raise' [Khalilov 1999: 139], Bezhta proper kikzi {кикзи} 'rearing' [Khalilov 1995: 146], which represent a borrowing from Avar xixi, inf. xixi-ze 'to feed / to rear, raise'.

    Reconstruction of Proto-Tsezic 'to eat (trans.)' is more problematic, because there are two equal candidates for this proto-meaning:

    1) East Tsezic *class=ũq ~ *class=ĩqʷ; lost in West Tsezic.

    2) West Tsezic *class=acʼ; lost in East Tsezic.

    There is no internal evidence for making a choice, but external North Caucasian comparison speaks in favour of *=ũq ~ *=ĩqʷ (its external cognates mean 'to eat' or 'to bite', whereas the more scarce cognates of =acʼ mean 'to drink').

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary verbal stem, meaning 'to eat (trans.)'.

    Reconstruction shape: Correspondences seem regular except for -ü- and the CVC-structure in Hinukh =iši ~ =üši.

    Semantics and structure: Primary verbal stem, meaning 'to eat (intrans.)'.

    Kwantlada Khwarshi: kʼimačʼ {кIимачI} 'egg' Karimova 2014; Khalilova 2009: 35, 57.

    1) *qõq-lV, retained in Hunzib, but lost in the rest of lects;

    2) cʼVmVcʼ ~ čʼVmVčʼ ~ qʼVmVcʼ ~ kʼVmVčʼ ~ kʼVnVčʼ in other languages. These forms are obvious related to each other, but phonetic correspondences are absolutely abnormal.

    The latter has the clear advantage in terms of distribution, but in fact its phonetic irregularity and absence of external North Caucasian etymology could suggest that we deal with a late introduction, which spread as an interdialectal borrowing. The exact source of cʼVmVcʼ ~ čʼVmVčʼ ~ qʼVmVcʼ ~ kʼVmVčʼ ~ kʼVnVčʼ is obscure, but it resembles certain Andian words for 'egg', cf. Chamalal proper čʼačʼã, pl. čʼačʼa-me, Gigatl Chamalal kʼučʼan pl. kʼučʼa-mi, Tindi kʼʸekʼʸama. The Tsezic forms could be a distortion of some Andian words with metathesis and various consonant assimilations/dissimilations.

    On the other hand, *qõq-lV, which is only documented for one peripheral language, Hunzib, possesses important external comparanda with the same meaning 'egg' that speaks in favour of its antiquity. Consequently, we fill the Proto-Tsezic slot with *qõq-lV.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root with the common l-suffix.

    Kwantlada Khwarshi: ezol [abs.] / ezala-s [gen.] {эзол} 'eye' [Karimova 2014; Khalilova 2009: 5, 28, 50, 61]. The archaic oblique stem ez- is attested in the locative adverb ez-e 'in the eye' [Khalilova 2009: 113]. In the plural form, with polysemy: 'eye / glasses'.

    Reconstruction shape: Correspondences seem regular except for secondary interactions between the direct and oblique stems in individual lects.

    Semantics and structure: Primary substantive root. Three oblique stems are reconstructible: *ħɔ- (as follows from East Tsezic data) and *ħɔ-zV-, *ħɔrV-zV- (as follows from West Tsezic data). It remains unclear which one reflects the original Proto-Tsezic paradigm. It should be noted that the oblique stem *ħɔrV-zV- has undergone metathesis > *ħɔzV-rV-, which has further spread into the direct stem in Dido and Khwarshi. Cf. similar rebuildings of the original paradigm in the words for 'bone', 'hand' q.v.

    Distinct from cǝcu {цǝцу} with polysemy: 'dissolved grease / bone marrow' [Isakov & Khalilov 2001: 173, 201; van den Berg 1995: 290]; in [Kibrik & Kodzasov 1990: 121], cǝcu is glossed as 'fat on meat'; in [Bokarev 1961: 168, 174], as generic 'fat'.

    Distinct from the more specific term mähä {маьгьаь} 'internal fat' [Khalilov 1995: 192, 306; Madieva 1965: 175].

    Distinct from mähä, which is glossed as generic 'fat' in [Kibrik & Kodzasov 1990: 121], but specified as 'internal fat' by M. Khalilov, p.c.

    Distinct from mäħä, which is glossed as generic 'fat' in [Kibrik & Kodzasov 1990: 121], but specified as 'internal fat' by M. Khalilov, p.c.

    Distinct from the more specific or borrowed terms, listed in [Khalilov & Isakov 2005: 441]:

    1) čečey {чечей} 'butter / fat' (apparently 'dissolved fat' is mentioned) [Khalilov & Isakov 2005: 390];

    2) qʼaƛu {къалIу} 'horse's fat' [Khalilov & Isakov 2005: 211];

    3) borrowed beʔenɬi {беэнлъи} 'fat / richness, fertility (of soil)' [Khalilov & Isakov 2005: 78] < Avar beʔenɬi 'fat / richness, fertility (of soil)';

    4) borrowed ƛʼaraɬi {кьаралъи} 'fatness / fat' [Khalilov & Isakov 2005: 224] < Avar ƛʼaraɬi 'fatness / fat'.

    Distinct from the more specific terms:

    1) qʼaƛi {къалIи} 'horse's fat / fat' ('fat' in general?) [Khalilov 1999: 145];

    2) riɬ {рилъ} 'butter / fat' (apparently 'dissolved fat' is meant) [Khalilov 1999: 218].

    Kwantlada Khwarshi: mu {му} 'fat' [Karimova 2014].

    In Bezhta, *mɔħV shifted to the specific meaning 'internal fat', having been superseded by *cǝcʷɨ (Hunzib cǝcu, Bezhta coco, Hinukh čečey), whose original meaning was probably 'dissolved fat'.

    In Sagada Dido, *mɔħV was superseded by *reɬː B 'butter' [NCED: 949].

    Replacements: {'butter' > 'fat'} (Sagada Dido).

    Reconstruction shape: Correspondences seem regular except for some vocalic peculiarities.

    Semantics and structure: Primary substantive root.

    Distinct from pudo {пудо} 'down (fine feathers)' [Khalilov 1995: 220, 331; Madieva 1965: 182].

    Distinct from pudo 'down (fine feathers)' [M. Khalilov, p.c.], despite the fact that pudo is glossed in [Kibrik & Kodzasov 1990: 45] as generic 'feather'.

    A second, probably more marginal, candidate is the inherited term lela {лела} with polysemy: 'wing / feather' [Khalilov 1999: 169, 356]; missing from [Kibrik & Kodzasov 1990].

    Distinct from borrowed ħuli {хIули} 'down' [Khalilov 1999: 268] < Avar ħuli 'down / feather'.

    Kwantlada Khwarshi: ħuli {хIули} 'feather' [Karimova 2014].

    In West Tsezic, the only documented inherited form is Mokok Dido lel 'wing / feather'; in other West Tsezic lects, it only means 'wing' < *ɬir(V) (~ *l-, -l-) [NCED: 762]. The Mokok polysemy is a transparent innovation. In the rest of West Tsezic lects, 'feather' is expressed by Georgian and Avar loanwords.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Kwantlada Khwarshi: cʼo {цIо} 'fire' [Karimova 2014; Khalilova 2009: 14].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Kwantlada Khwarshi: čuʕa {чугIа} 'fish' [Karimova 2014; Khalilova 2009: 15].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root. The oblique stem is *bɨsːʷǝ-ro- A, as it can be reconstructed at least for Proto-West Tsezic (where the oblique stem has supplanted the direct one).

    Another verb, which can be used in the same meaning, is kʼoƛʼ(e) 'to jump' [Khalilov & Isakov 2005: 235] (thus, with polysemy: 'to jump / to fly'). E.g., in combination with hawa 'air': hawa-ƛʼo kʼoƛʼ(e) 'to fly', literally 'to jump in the air' [Khalilov & Isakov 2005: 132].

    Cf. also the verb noχ 'to come / to fly here' q.v.

    Distinct from boržizi class=iq 'to fly', applicable to aircraft, literally 'to become boržizi' [Khalilov & Isakov 2005: 99], borrowed from Avar b=orž-ize 'to fly'.

    Cf. also the verb class=oχi {боха} 'to run away, go away' [Khalilov 1999: 69], used for the meaning 'to fly away' as in "The fledgeling, fly (away)!" [Khalilov 1999: 296], and the verb neχ 'to come / to fly here' q.v.

    Additionally, the collocation class=etʼu-n kʼoƛi {бетIункIолIа} with polysemy: 'to jump / to fly' is quoted in [Khalilov 1999: 48], where =etʼu-n is the past tense form of class=etʼu 'to tear off (intrans.)' [Khalilov 1999: 47] and the inflected verb kʼoƛi {кIолIа} 'to run' [Khalilov 1999: 164]. Apparently =etʼu in the collocation for 'to fly' has nothing to do with Tindi etʷ- 'to fly' and cognate verbs in other Andian languages.

    Kwantlada Khwarshi: etʷ {этва} 'to fly' [Karimova 2014; Khalilova 2009: 18, 181]. Cf. the example: "Now this boy turned into a pigeon, this boy flew away" [Khalilova 2009: 147].

    In West Tsezic lects (Hinukh, Kidero Dido) 'to fly' is expressed by various verbs for 'to go' and 'to come', which differ between languages. Theoretically, one of these verbs for 'to go', namely *class=ẽƛʼ- or *class=ẽkʼ-, can be reconstructed with the additional meaning 'to fly' at the Proto-Tsezic level, but it is more likely that we deal with a late innovation 'to go' > 'to fly' in Hinukh, Kidero Dido.

    In Sagada Dido and Khwarshi, inherited forms are superseded by Avar and Tindi loanwords.

    Replacements: {'to go' > 'to fly'} (Hinukh, Kidero Dido).

    Reconstruction shape: Occasional dissimilative deglottalization kʼ > g in Bezhta proper, irregular cʼ for expected čʼ in Tlyadal Bezhta (or, if the protoform was *class=Vkʼicʼ-, irregular čʼ for expected cʼ in Hunzib).

    Semantics and structure: Primary verbal stem; polysemy 'to jump / to fly' can be reconstructed at least for Proto-East Tsezic.

    Distinct from čʼumal {чIумал}, whose exact meaning is, however, uncertain: glossed as 'shin, leg' [Isakov & Khalilov 2001: 180], 'leg from knee till foot' in [van den Berg 1995: 293], generic 'leg' in [Kibrik & Kodzasov 1990: 30]; Russian 'нога', which can mean 'foot', 'leg' or 'foot + leg' in [Bokarev 1961: 169].

    Distinct from specific halo {гьало} 'foot' [Khalilov 1995: 73], which is apparently rare and obsolete.

    Distinct from čʼamal {чIамал} with polysemy: 'shin / leg / stalk, stem' [Khalilov 1995: 279; Madieva 1965: 194].

    Distinct from čämäl (a typo for čʼämäl?), which is quoted in [Kibrik & Kodzasov 1990: 30] as a generic term ('foot + leg'), but actually seems to specifically denote 'leg'.

    Distinct from čʼämäl, which is quoted in [Kibrik & Kodzasov 1990: 30] as a generic term ('foot + leg'), but actually seems to specifically denote 'leg'.

    Distinct from kʼoncʼu {кIонцIу} 'leg' [Khalilov & Isakov 2005: 236; Kibrik & Kodzasov 1990: 30], although in [Khalilov & Isakov 2005] this word is glossed as 'нога (вся)', i.e., 'leg + foot'(?).

    Distinct from other words, quoted in [Khalilov & Isakov 2005] with the non-specified gloss 'нога'.

    1) oʁ ~ oˤʁ {огъ, оIгъ} [Khalilov & Isakov 2005: 283] with polysemy 'leg / thigh, hip / trouser leg' (apparently not 'foot').

    2) hobo {гьобо} 'foot' [Khalilov & Isakov 2005: 144], the specific meaning 'foot' is seen from the illustrative examples: “to kick with the foot”, “to hit with the foot”.

    Distinct from bula {була} 'leg of animal / hoof' [Khalilov & Isakov 2005: 107].

    Distinct from kʼončʼu {кIончIу} with polysemy: 'leg / furniture leg' [Khalilov 1999: 165; Kibrik & Kodzasov 1990: 30]; it should be noted that in [Kibrik & Kodzasov 1990], this word is glossed as generic 'foot + leg'.

    Distinct from several more marginal or specific terms:

    1) ʡoʁ {гIогъ, гIоIгъ} (Asakh oˤʁ) [Khalilov 1999: 110] with polysemy: 'leg / thigh, hip' (apparently not 'foot');

    2) ʁol {гъол} 'foot' or 'leg', glossed as Russian 'нога' [Khalilov 1999: 92], rarely used;

    3) χotʼo {хотIо} with polysemy: 'foot / step / furniture leg' [Khalilov 1999: 249], cf. the compound χotʼo-reƛʼa 'human extremities', literally 'foot + hand/arm' [Khalilov 1999: 250];

    4) qˤošo {хъоIшо} 'heel (of human) / leg (of animal)' [Khalilov 1999: 257];

    5) buˤla ~ bula {буIла, була} 'leg (of animal) / hoof' [Khalilov 1999: 73, 76].

    Distinct from kʼäkʼä 'leg' [Bokarev 1959: 144].

    Kwantlada Khwarshi: lolo ~ lʸolʸo {лоло} with polysemy: 'foot / leg' [Karimova 2014; Khalilova 2009: 20, 75, 77, 119, 123]. Distinct from the more specific term kʼaˤkʼa ~ kʼakʼa 'leg' [Khalilova 2009: 17, 46, 80].

    In all Bezhta dialects it was superseded by *χɔbɔ A [NCED: 454] with polysemy 'foot / leg'. The original meaning of *χɔbɔ is unclear, the only Hinukh cognate means 'foot', but it is not the basic Hinukh term for this meaning. In Bezhta proper, *ɬɔlǝ is still documented as an obsolete word for 'foot'.

    In Sagada Dido, the default term for 'foot' seems to be χotʼo - a form of unclear origin, it is also attested in Kidero Dido with the same meaning 'foot', but it is not the basic Kidero term for this meaning.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root, meaning specifically 'foot'.

    Kwantlada Khwarshi: class=ecʼː-u {лецIцIу} 'full' [Karimova 2014; Khalilova 2009: 21, 168].

    Reconstruction shape: The vocalic correspondence between Proto-East and Proto-West forms is irregular.

    Semantics and structure: *class=ǝcʼ- is the verb 'to be(come) full'; the adjectival meaning 'full' in modern lects is expressed by various synchronic participles.

    The Hinukh verb for 'to give' is suppletive in respect of recipient person: neƛ is used with 1^st/2^nd p. of the recipient 'to give to me, us, you' / toƛ is used with 3^rd p. of the recipient 'to give to him, her, it, them' [Forker 2013: 484 f.]. It must be noted that the distribution is not absolutely strict; browsing through [Forker 2013] provides some aberrant examples among a great number of regular instances: cf. neƛ in the constructions 'he to them' [Forker 2013: 319, 395], 'I to him' [Forker 2013: 250].

    We treat neƛ and toƛ as synonyms, although historically they represent paronymous formations n=eƛ and t=oƛ with fossilized directional prefixes.

    The Kidero Dido verb for 'to give' is suppletive in respect to the recipient person: neƛ is used with 1^st/2^nd p. of the recipient 'to give to me, us, you' / teƛ is used with 3^rd p. of the recipient 'to give to him, her, it, them'. We treat neƛ and teƛ as synonyms, although historically they represent paronymous formations n=eƛ and t=eƛ with fossilized directional prefixes.

    Kwantlada Khwarshi: iƛ {илIа} 'to give' [Karimova 2014; Khalilova 2009: 327] and tɨƛ ~ tuƛ 'to give' [Khalilova 2009: 327].

    Reconstruction shape: Correspondences seem regular except for the strange alternation *i in the plain form vs. *ɨ in the prefixed forms.

    Semantics and structure: Primary verb. The plain root *ʔiƛ- is only retained in Khwarshi. Normally we find the prefixed stems *n=ɨƛ- and *t=ɨƛ- (in East Tsezic, only *n=ɨƛ- is used).

    In West Tsezic, the usual situation is the grammatical opposition between two stems in respect to the recipient person: stem-A with 1^st/2^nd p. vs. stem-B with 3^rd p. In Hinukh-Dido, this opposition looks as follows: *n=ɨƛ- / *t=ɨƛ-. Somewhat differently in Khwarshi: *ʔiƛ- / *t=ɨƛ-. Because of such a material difference, it is hard to project the described grammatical opposition onto the Proto-West Tsezic level. More likely, we deal with a later grammatical introduction that has affected neighboring West Tsezic lects.

    We prefer to treat *n= and *t= as old directional prefixes (a normal pattern for the verb for 'to give'), which probably got desemanticized and fossilized already in Proto-Tsezic. The competing reconstruction of *n= and *t= as unique recipient person exponents does not find additional confirmation.

    Distinct from adverbial aχ-o {ахо} 'good, in the proper way / very' [Khalilov 1995: 30; Madieva 1965: 145, 148].

    Distinct from adverbial aχ-o {ахо} 'good, in the proper way / very' (Ya. Testelets, p.c.).

    Distinct from two more marginal words:

    1) zapʼanaw {запIанав} with polysemy: 'beautiful / good / lucky / strong' [Khalilov & Isakov 2005: 171], looks like an Avarism, but the source form has not been found;

    2) lazːataw {лаззатав} 'pleasant, good, convenient' [Khalilov & Isakov 2005: 239] < Avar lazata-w 'pleasant / tasty'.

    Distinct from less common words:

    1) zapʼana {запIана} 'beautiful / good' [Khalilov 1999: 121] (a loan?);

    2) lazatab {лазатаб} 'pleasant, good' [Khalilov 1999: 167] < Avar lazata-b 'pleasant / tasty';

    3) cʼikʼaraw {цIикIарав} 'good / superfluous / numerous' [Khalilov 1999: 277] < Avar cʼikʼːara-w 'elder / magnificent, outstanding'.

    Kwantlada Khwarshi: class=og-u {огу} 'good' [Karimova 2014; Khalilova 2009: 5, 100]. Widely applicable.

    1) East Tsezic *kʼotʼV [NCED: 724]; lost in West Tsezic;

    2) West Tsezic *class=ɨgV (~ -ǝ-) [NCED: 643]; lost in East Tsezic.

    There is no internal Tsezic evidence that would allow us to make a single choice. External comparison, however, clearly suggests that the Proto-Tsezic term for 'good' was *class=ɨgV, its cognates mean 'good' or 'right' in Nakh, Avaro-Andian, Dargi, Lezgian. On the other hand, Tsezic *kʼotʼV is either unetymologizable or originates from the meaning 'soft'.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary adjectival stem.

    A second term for 'green' is heƛ-aɬ-co-ʔos {гьелIалъцоъос} [Khalilov 1995: 80], derived from the noun heƛ-aɬ-co 'green color', further to heƛe 'walnut (a fruit)'. Cf. the attested example: “green dye (paint)” [Khalilov 1995: 281]. This adjective is apparently more marginal than nič-d-iy-o, because heƛ-aɬ-co-ʔos is not quoted in [Madieva 1965].

    A second candidate is nik-d-iy-u, glossed in [Kibrik & Kodzasov 1990: 234, 235] with polysemy: 'blue / green', but other sources only provide the meaning 'blue' for this adjective: [Khalilov & Isakov 2005: 277; Forker 2013: 172].

    Distinct from the loanword ʡurčinaw {гIурчинав} 'green (applied to plants)' [Khalilov & Isakov 2005: 158] < Avar ʕurčːina-w 'green (of plants)'.

    Kwantlada Khwarshi: the same loanword qayƛe {хъайлIе} 'green' [Karimova 2014].

    In West Tsezic, expressions for 'green' normally represent various denominative stems (the attested starting nouns are 'first spring grass' and 'leaf'). Such adjectives should be regarded as relatively recent formations.

    In Inkhokwari Khwarshi, an Andian loanword is attested.

    Replacements: {'first spring grass' > 'green'} (Hinukh); {'leaf' > 'green'} (Khwarshi proper).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Probably a primary adjectival root.

    A second candidate is müčʼ 'head hair', quoted in [Khalilov 1995], but missing from [Madieva 1965]. We treat kẽyã and müčʼ as synonyms.

    Distinct from kera 'a single hair' [Kibrik & Kodzasov 1990: 42].

    Distinct from kerä 'a single hair' [Kibrik & Kodzasov 1990: 42].

    Distinct from the more marginal borrowed term kočori {кочори} 'head hair / forelock' [Khalilov & Isakov 2005: 206] < Georgian kočori 'topknot, tuft of hair'.

    Distinct from the specific term čud {чуд} 'mane / forelock, long head hair' [Khalilov & Isakov 2005: 392], which is, however, quoted in [Kibrik & Kodzasov 1990: 42] as the generic term for 'head hair'.

    Distinct from specific peħ {пехI} 'pubic hair' [Khalilov & Isakov 2005: 292].

    Distinct from mus {мус} 'a single hair' [Khalilov 1999: 194, 311; Kibrik & Kodzasov 1990: 42].

    Distinct from mus {мус} with polysemy: 'a single hair (of human or animal) / body hair (of human) / fur (of animal)' [Abdulaev 2014].

    Distinct from ka 'a single hair' [Sharafutdinova & Levina 1961: 113; Bokarev 1959: 147, 152, 156].

    Distinct from Inkhokwari kõ ~ ko 'a single hair' [Kibrik & Kodzasov 1990: 42; Bokarev 1959: 147, 148, 150, 152].

    Kwantlada Khwarshi: kode {коде} [Karimova 2014; Khalilova 2009: 34]. It denotes collective 'hair'. Cf. the attested examples: "The girl cut (her) hair by accident" [Khalilova 2009: 324], "As soon as the girl fell asleep, the neighbor tied her hair to the tree" [Khalilova 2009: 87, 403].

    There are also two other Kwantlada terms: ko 'hair' [Khalilova 2009: 57, 263] and mɨs ~ mus 'hair' [Khalilova 2009: 64, 65, 135]. Their exact meanings are unknown, at least ko is expected to mean 'a single hair'. Cf. the found example for mɨs ~ mus: "When the man sat on the top of the yoke bending, (his) hair had fallen" [Khalilova 2009: 135].

    NCED: 805. Distribution: The relevant Tsezic forms can be summarized in the following table:

'HAIR'	Hunzib	Bezhta proper	Bezhta (other)	Hinukh	Dido	Khwarshi proper	Inkhokwari Khwarshi
*kẽ B [NCED: 697]	single hair / head hair	single hair / head hair	single hair			single hair	single hair
*mučʼ B [NCED: 805]		head hair	head hair	neck	neck	neck	neck
*mosː (~ -u-, -s) [NCED: 805]				single hair / head hair (pl.)	single hair / body hair	hair (unspecified)	hair (unspecified)
*kɔdV A [NCED: 705]					head hair	čoda head hair	head hair
*čo A [NCED: 347]		čoy hair in horse's tail		čud long head hair			ču mane

    The Proto-East Tsezic system can be reconstructed as *kẽ 'a single hair' / *mučʼ 'hair (collective), head hair'. In Hunzib, the system became simplified: *mučʼ was eliminated and *kẽ acquired the polysemy 'head hair / body hair / a single hair'. In Bezhta proper, *kẽ acquired the polysemy 'head hair / a single hair' (perhaps independently from the same process in Hunzib).

    The Proto-West Tsezic system can be reconstructed as *kẽ 'a single hair' / *kɔdV 'head hair'. In Hinukh & Dido, *kẽ 'a single hair' got lost, having been superseded by *mosː (the original semantics of *mosː was 'a kind of hair'; more exact specification is difficult, but the proto-meaning 'body hair, fur' is very probable, cf. its meaning in Dido); such a replacement should be treated as contact-driven homoplasy between Hinukh & Dido. Additionally, in modern Hinukh, 'head hair' is secondarily expressed as the pl. form of *mosː. In Khwarshi proper, *kɔdV 'head hair' was phonetically influenced by *čo 'horsehair', having substituted initial k- for č- (the similar hybrid form čud is observed in Hinukh, but its meaning 'long head hair' suggests that we more probably deal with a reverse process, when *čo 'mane' > 'long head hair' was influenced by *kɔdV 'head hair').

    External comparison (first of all, the Andian comparanda) suggests that the East Tsezic system must be projected onto the Proto-Tsezic level: *kẽ 'a single hair' / *mučʼ 'head hair'. In Proto-West Tsezic, *mučʼ 'head hair' shifted to the meaning 'neck', having been superseded by *kɔdV in the meaning 'head hair'. The original Proto-Tsezic semantics of *kɔdV is unclear; its North Caucasian etymology (if correct!) suggests the meaning 'bush, crown of a tree'.

    Replacements: {'a single hair' > 'head hair'} (Hunzib, Bezhta proper, Hinukh); {'head hair' > 'neck'} (Proto-West Tsezic).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Distinct from bɨcu {быцу} 'arm' [Isakov & Khalilov 2001: 48; Kibrik & Kodzasov 1990: 26; van den Berg 1995: 288; Bokarev 1961: 152, 180].

    Distinct from specific rɨƛʼa {рыкьа} with polysemy: 'sleeve / front leg of animal' [Isakov & Khalilov 2001: 145; van den Berg 1995: 331].

    Distinct from bico {бицо} 'arm' [Khalilov 1995: 48, 333; Madieva 1965: 151].

    Distinct from specific yiƛʼa {йикьа} 'sleeve' [Khalilov 1995: 121].

    Distinct from bico 'arm' [Kibrik & Kodzasov 1990: 26].

    Distinct from specific riƛʼa with polysemy: 'sleeve / foreleg' [Kibrik & Kodzasov 1990: 26, 130].

    As noted in [Kibrik & Kodzasov 1990: 26], there also exists a Khoshar-Khota generic term for 'hand + arm': the compound bico-riƛʼa, literally 'arm' + 'sleeve'. Statistical difference between the more specific koro and the more generic bico-riƛʼa is not described, but we may assume that the situation is the same as in Tlyadal Bezhta q.v.

    Distinct from bico 'arm' [Kibrik & Kodzasov 1990: 26].

    Distinct from specific riƛʼa 'sleeve' [Kibrik & Kodzasov 1990: 130].

    According to [Kibrik & Kodzasov 1990: 26], there also exists a Tlyadal generic term for 'hand + arm': the compound bico-riƛʼa, literally 'arm' + 'sleeve'. This one, however, seems more marginal in the meaning 'hand' than the more specific koro, because in all the found examples only koro is attested: “The dog licks the hand” [Kibrik & Testelets 2004: 230], “Fire burnt the boy's hand” [Kibrik & Testelets 2004: 231], “The boy burnt his hand with fire” [Kibrik & Testelets 2004: 231], “Until it (a snake) bit him, his hand was healthy” [Kibrik & Testelets 2004: 266], “His hands are long” [Kibrik & Testelets 2004: 280].

    Cf. the specific adverb kʷaː {ква̄} 'in the hand(s)' [Khalilov & Isakov 2005: 197].

    Distinct from mecu {мецу} with polysemy: 'arm / sleeve / draft bar' [Khalilov & Isakov 2005: 265; Kibrik & Kodzasov 1990: 26].

    Distinct from qot {хъот} with polysemy: 'paw / palm of hand / hand / palm slap' [Khalilov & Isakov 2005: 363].

    Distinct from mecu {мецу} with polysemy: 'arm / draft bar' [Khalilov 1999: 188; Kibrik & Kodzasov 1990: 26].

    Distinct from geša 'arm' [NCED: 448].

    Distinct from geša 'arm' [Kibrik & Kodzasov 1990: 26; Bokarev 1959: 150].

    Kwantlada Khwarshi: lɨƛʼa ~ luƛʼa ~ liƛʼa {лыкьа, лукьа} with polysemy: 'hand / arm' [Karimova 2014; Khalilova 2009: 45, 67, 79, 113]. The variants lɨƛʼa ~ luƛʼa are from [Karimova 2014]. In [Khalilova 2009], it is normally transcribed as liƛʼa and seldom as lɨƛʼa. Numerous textual instances for the meaning 'hand' are available in [Khalilova 2009], but no examples for 'arm' are known, however.

    Distinct from Kwantlada qot 'palm of hand' [Khalilova 2009: 14, 32].

    Distinct from geša 'shoulder' [Khalilova 2009: 45].

    NCED: 706. Distribution: The relevant forms can be summarized in the following table:

'HAND'	Hunzib	Bezhta	Hinukh	Dido	Khwarshi (proper, Inkhokwari)	Kwantlada Khwarshi
*kʷɨrV [NCED: 706]	hand	hand	hand
*bɨ̃cu A [NCED: 307]	arm	arm	arm / sleeve	arm
*rɨƛʼa A [NCED: 779]	sleeve / foreleg	sleeve / foreleg		hand	hand	hand / arm
*gɔšːa (~ -e-, -š-) [NCED: 448]					arm	shoulder

    Proto-Tsezic *bɨ̃cu can be safely reconstructed with the meaning 'arm'; in Khwarshi, it was superseded by *gɔšːa, whose original meaning is unclear (its Kwantlada Khwarshi meaning and external etymology, if correct, could point to something like 'shoulder' [NCED: 448]).

    Reconstruction of the Proto-Tsezic term for 'hand' is somewhat more problematic. Two stems enter into competition:

    1) *kʷɨrV, which can be posited at least as Proto-East Tsezic 'hand' (plus 'hand' in Hinukh, lost in the rest of West Tsezic lects);

    2) *rɨƛʼa, which is attested as 'hand' in West Tsezic: Dido plus Khwarshi minus Hinukh. In East Tsezic, it means 'sleeve / foreleg'.

    We posit *kʷɨrV as the Proto-Tsezic term for 'hand'. In the Dido-Khwarshi cluster it was superseded by *rɨƛʼa.

    Theoretically, however, the opposite solution is also possible: *rɨƛʼa was the Proto-Tsezic term for 'hand', which was superseded by *kʷɨrV in Proto-East Tsezic and secondarily in Hinukh under the influence on the part of East Tsezic.

    External comparison speaks in favour of Proto-Tsezic *kʷɨrV 'hand' (its cognates mean 'hand' in Avar, Nakh, Lak, Lezgian, Khinalug [NCED: 706]). But, strictly speaking, we are in a situation of semantic criss-crossing, since external cognates of *rɨƛʼa mean 'hand' in Andian [NCED: 779].

    From the geographical point of view, Proto-Andian *riƛʼa 'hand' / Dido-Khwarshi *rɨƛʼa 'hand' is an areal isogloss, which affects the Andian languages together with neighboring Dido & Khwarshi. Thus we suppose that *riƛʼa was the Proto-Andian term for 'hand' (an Andian introduction), which has later influenced the adjacent Tsezic lects (namely Dido & Khwarshi with all their dialects). The original Proto-Tsezic meaning of *rɨƛʼa is unknown (cf. its meanings 'sleeve / foreleg' in East Tsezic).

    Reconstruction shape: Correspondences seem regular, except for the consonant metathesis in the oblique stem.

    Semantics and structure: Primary substantive root, meaning specifically 'hand'. Two oblique stems can be reconstructed. The first one is *kʷṼ-, its antiquity is beyond doubt since it is retained in East Tsezic and adverbially in Hinukh. The second oblique stem is *kʷɨrV-zV-, metathesized to *kʷɨzVra-; this one is reflected in the Hinukh paradigm kʷezey [abs.] / kʷeze-ra- [obl.], where the abs. stem is a back formation. Apparently the oblique stem *kʷɨrV-zV- is a secondary Proto-West Tsezic or even Proto-Hinukh introduction, cf. the same pattern in the words for 'bone', 'eye' q.v.

    Distinct from oqʼru {окъру} with polysemy: 'skull, cranium / head, noddle' [Khalilov & Isakov 2005: 283].

    Kwantlada Khwarshi: qʼˤem {къеIм} 'head' [Karimova 2014; Khalilova 2009: 16].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Distinct from the more marginal verb tuq {тухъа}, which is glossed as 'to hear' in [Isakov & Khalilov 2001: 155, 233] (with only one example: “He has heard the news”) and as 'to listen' in [van den Berg 1995: 336].

    Distinct from the analytic expressions for 'to listen': ãʁa-tuq-le, ãʁa-tuq-kʼe {ангъатухъла, ангъатухъкIа} [Isakov & Khalilov 2001: 25, 233; van den Berg 1995: 283], literally 'to put the ear on' with ãʁa 'ear' q.v. and tuq-le, tuq-kʼe 'to lean on, put on'. Additionally, the compound ãʁa-reʁ {ангъарегъа} 'listen' in quoted in [Isakov & Khalilov 2001: 25, 233] with reʁ '?'.

    Distinct from enekzi class=aq {энекзийахъал} with polysemy: 'to listen / to obey' [Khalilov 1995: 292, 336], literally 'to become enekzi', the latter is borrowed from Avar ʕenekːize 'to listen'. See [Madieva 1965: 114] for this complex verb pattern, typical of loanwords.

    Distinct from čoq {чохъал} 'to hear of, be aware of' [Khalilov 1995: 277], čʼaɬ {чIалъал} 'to hear that, be aware of' [Khalilov 1995: 279].

    Distinct from ä̃ʁä-kʼ-äː {ъʻаI^нгъʻаIкIа̄ьл} 'to hear out / to overhear' [Khalilov 1995: 290], derived from ä̃ʁä 'ear' q.v. with the denominative (factitive, causative) suffix -kʼ [Kibrik & Testelets 2004: 270; Madieva 1965: 113] and iterative -äː.

    Distinct from äʁä-kʼ-d-äː 'to listen' [Kibrik & Kodzasov 1988: 172], , derived from äʁä 'ear' q.v. with the denominative (factitive) suffix -kʼ and the durative suffix -da [Kibrik & Testelets 2004: 270; Madieva 1965: 113] plus iterative -äː.

    Distinct from ʕenekzi class=aq 'to listen' [Kibrik & Kodzasov 1988: 172], literally 'to become ʕenekzi', the latter is borrowed from Avar ʕenekːize 'to listen'.

    Distinct from ʡenekezi class=iq ~ ʡenekizi class=iq {гIенекези, гIенекизи бихъа} with polysemy: 'to listen / to obey' [Khalilov & Isakov 2005: 155], literally 'to become ʡenekezi', the latter is borrowed from Avar ʕenekːize 'to listen'.

    Distinct from the expressions for 'to listen':

    1) teq-er {техъра} with polysemy: 'to listen / overhear' [Khalilov 1999: 235, 376], causative from teq 'to hear';

    2) ʡenekizi class=oq {гIенекизи охъа} with polysemy: 'to listen / to obey' [Khalilov 1999: 108, 376], literally 'to become ʡenekizi', the latter is borrowed from Avar ʕenekːize 'to listen'.

    Kwantlada Khwarshi: tuq {тухъа} 'to hear' [Karimova 2014; Khalilova 2009: 77, 82, 84, 190, 305]. Polysemy: 'to hear / to listen'. The model for the meaning 'to hear': absolutive (sound) + lative (recipient); the model for the meaning 'to listen': ergative (recipient) + absolutive (sound) [Khalilova 2009: 305].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary verbal stem, meaning 'to hear' with the model "absolutive (sound) + dative (recipient)".

    Kwantlada Khwarshi: lokʼo {локIо} 'heart' Karimova 2014; Khalilova 2009: 45.

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Kwantlada Khwarshi: šelʸu {шелʼу} 'horn' [Karimova 2014; Khalilova 2009: 7, 20, 59].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Kwantlada Khwarshi: do [abs.] / de [erg.] / di- [obl.] 'I' [Karimova 2014; Khalilova 2009: 142].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary pronominal root. The oblique stem is *di-.

    Distinct from class=uh-kʼe {бугькIа} 'to spoil / to kill', class=uh-le {бугьла} 'to spoil / to destroy / to kill' [Isakov & Khalilov 2001: 45, 239; van den Berg 1995: 337; Bokarev 1961: 166, 181] - causatives from class=uhu 'to die / to get spoilt' q.v.

    Distinct from the causatives from the more rare verbs for 'to die': ɬiyi-r {лъийра} with polysemy: 'to finish (trans.) / to kill' [Khalilov 1999: 175], ɬiqu-r {лъихъура} with polysemy: 'to dry (trans.) / to kill' [Khalilov 1999: 175], ʁutʼ-er {гъутIра} 'to kill' (only of animals?) [Khalilov 1999: 94].

    A second equivalent for 'to kill' is class=iha-x {игьаха} 'to kill' [Sharafutdinova & Levina 1961: 115] with the example "The wolf said: 'Should I kill you?'" [Sharafutdinova & Levina 1961: 117] - regular causative from class=ih 'to die' q.v.

    Distinct from the more specific verb class=iχʷ-ad 'to slaughter' [Sharafutdinova & Levina 1961: 119; Bokarev 1959: 148].

    Distinct from the more specific Inkhokwari verb class=uχ-ad 'to slaughter' [Bokarev 1959: 148].

    Kwantlada Khwarshi: class=uwoχ {увоха} 'to kill' [Karimova 2014; Khalilova 2009: 70]. The contracted variant class=oːχ is also attested [Khalilova 2009: 191]. Cf. some examples: "Having fallen, the post killed the man" [Khalilova 2009: 70]; "When they were killed in Finland, and when there was a treaty with them, ..." [Khalilova 2009: 191]; "He was going to kill her" [Khalilova 2009: 195]; "If you like, kill me" [Khalilova 2009: 414]; "Magomed killed Mesedo at the place where she was sleeping" [Khalilova 2009: 420]; "his raven had been killed" [Khalilova 2009: 422]; and so on.

    A second Kwantlada candidate is the verb aˤh (i.e., class=aˤh?), quoted with the gloss 'to kill' in [Khalilova 2009: 17], but no examples have been found.

    Distinct from the more specific Kwantlada verb class=uχ-ad ~ class=uχˤ-ad 'to slaughter' [Khalilova 2009: 307, 315, 393, 423].

    In the majority of West Tsezic lects plus Khoshar-Khota Bezhta, the meaning 'to kill' is expressed by synchronic causative stems from the verb 'to die' q.v. (*class=iχʷV [NCED: 635]). In Hunzib, it coexists with *class=iƛʼV.

    In Khwarshi, 'to kill' is derived from the verb 'to slaughter' (*class=ĩχːʷV A [NCED: 635], although the causative stem from 'to die' is also present.

    External comparison definitely suggests that the Proto-Tsezic verb for 'to kill' was *class=iƛʼV, whereas causative stems from 'to die' represent late introductions (formally, such a causative pattern can be reconstructed for the Proto-West Tsezic level, but it is clearly secondary in Khoshar-Khota Bezhta & Hunzib).

    Replacements: {'to die' > 'to kill'} (Hunzib, Khoshar-Khota Bezhta, Hinukh, Dido); {'to slaughter' > 'to kill'} (Khwarshi).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary verbal stem.

    In [NCED: 594], Hunzib ãga ‘knee-cap’ is quoted, not found in other sources.

    Distinct from the more specific term ãga {а^нга} 'knee-cap, patella' [Khalilov 1995: 34], which is quoted, however, in [Madieva 1965: 148] as the basic equivalent for 'knee'.

    Distinct from micna 'shin (part of leg from knee to foot)' [Kibrik & Kodzasov 1990: 31].

    Distinct from mičnä 'shin (part of leg from knee to foot)' [Kibrik & Kodzasov 1990: 31].

    Distinct from the more specific qʼontu {къонту} 'knee-cap, patella' [Khalilov & Isakov 2005: 220]. In [Kibrik & Kodzasov 1990: 32], erroneously glossed as generic 'knee'.

    Distinct from specific qʼontu {къонту} 'knee-cap, patella' [Khalilov 1999: 152], which is transcribed as qʼˤontu and erroneously glossed as generic 'knee' in [Kibrik & Kodzasov 1990: 32].

    A second Inkhokwari term for 'knee' is ƛʼehemu {кьегьему} [Karimova 2014; Bokarev 1959: 145, 150]; according to [Karimova 2014], gurtu and ƛʼehemu are full synonyms, but the latter one is less frequently used.

    Kwantlada Khwarshi: gurtu {гурту} 'knee' [Karimova 2014; Khalilova 2009: 14, 46].

    Nevertheless, *bičnV [NCED: 1042] can be safely reconstructed as Proto-Tsezic 'knee'. It is retained in its basic meaning in some East Tsezic (Hunzib, Bezhta proper) and some West Tsezic (Hinukh, Kidero Dido) lects.

    In Khoshar-Khota Bezhta & Tlyadal Bezhta, *bičnV was superseded by *ʔãgV, having shifted in the meaning 'shin (part of leg from knee to foot)'. The original meaning of *ʔãgV [NCED: 594] is unclear, but it could be the Proto-East Tsezic term for 'knee-cap' (it means this in Bezhta proper).

    In Sagada Dido and Khwarshi proper, the basic term for 'knee' was superseded by *qʼˤɔ(n)tV (if the field records are correct). The original Proto-West Tsezic meaning of *qʼˤɔ(n)tV was 'knee-cap', as it is retained in Hinukh and non-Sagada Dido. In any case, *qʼˤɔ(n)tV with the generic meaning 'knee' is a secondary match between Sagada Dido and Khwarshi proper.

    In Inkhokwari Khwarshi, the basic term for 'knee' was superseded by two unclear forms: gurtu (cf. [NCED: 434]) and ƛʼehemu.

    Replacements: {'knee' > 'shin (part of leg from knee to foot)'} (Khoshar-Khota Bezhta, Tlyadal Bezhta).

    Reconstruction shape: Correspondences seem regular except for the nasal assimilation b-n > m-n in East Tsezic.

    Semantics and structure: Primary substantive root.

    Distinct from čʼaɬ {чIальа} 'to get to know' [Khalilov 1999: 285].

    Kwantlada Khwarshi: class=iqʼ {ликъа} 'to know' [Karimova 2014; Khalilova 2009: 27, 152, 181]. Cf. the present form class=iqʼ-še [Khalilova 2009: 152].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary verbal root.

    Distinct from the more specific term ƛʼapa {кьапа} 'large leaf' [Isakov & Khalilov 2001: 101, 210].

    Distinct from ƛʼapa {кьапа}, which is glossed as 'leaf (of tree or plant)' [Khalilov 1995: 161, 314]; not quoted in [Madieva 1965]. Apparently ƛʼapa is a more marginal term than ƛibo, because all the attested examples for the meaning 'leaf' contain the word ƛibo: “to rustle the leaves” [Khalilov 1995: 37], “Leaves are stirring” [Khalilov 1995: 176], “Stirring of leaves” [Khalilov 1995: 176], “tobacco leaf” [Khalilov 1995: 235], “maple leaves” [Khalilov 1995: 283], “There are a lot of leaves in the tree” Madieva 1965: 89], “Leaves have fallen from the tree” [Madieva 1965: 89].

    Kwantlada Khwarshi: ƛib {лIиб} 'leaf' [Karimova 2014; Khalilova 2009: 14, 202, 479].

    Reconstruction shape: Correspondences seem regular except for a in Kidero Dido and loss of final *-u in some West Tsezic lects.

    Semantics and structure: Primary substantive root.

   Two synonymous expressions for 'to lie' are quoted in [Khalilov 1995]: primary =utʼ and complex =ãko =eƛʼe. Both with polysemy: 'to lie / to lie down'.

    Two synonymous expressions for 'to lie down' are quoted in [Kibrik & Kodzasov 1988]: primary =utʼ and complex =ako =eƛʼe. We assume that both actually possess the polysemy: 'to lie / to lie down'; the former one, =utʼ, possesses the full polysemy: 'to lie / to lie down / to sleep'.

    Two synonymous expressions for 'to lie down' are quoted in [Kibrik & Kodzasov 1988]: primary =utʼ and complex =ãko =eƛʼe. We assume that both actually possess the polysemy: 'to lie / to lie down'; the former one, =utʼ, possesses the full polysemy: 'to lie / to lie down / to sleep'.

    Two synonymous expressions for 'to lie' are quoted in [Khalilov 1999] and illustrated with a number of textual examples: =aˤƛʼu (also means 'to fall') and kec (also means 'to sleep'). We have to treat them as synonyms.

    Kwantlada Khwarshi: class=aqˤ {аIхъаI} with polysemy: 'to lie / to lie down' [Karimova 2014; Khalilova 2009: 16, 180].

    NCED: 1035. Distribution: Tsezic verbs for 'to lie' and 'to sleep' can be summarized in the following table:

'LIE', 'SLEEP'	Hunzib	Bezhta	Hinukh	Kidero Dido	Sagada Dido	Khwarshi
*class=itʼʷ [NCED: 1035]	lie / sleep	lie / sleep	lie / sleep
*class=ãkV [NCED: 644]		lie
*class=aˤƛʼV				lie / fall
*kic				lie / sleep	sleep
class=aqˤ(V)						lie / sleep
?					'be in the lateral position' = lie
?		class=üčäːχe 'sleep'
?						ƛVs 'sleep'

    The verb *class=itʼʷ can be safely postulated as the Proto-East Tsezic basic term with polysemy 'to lie / to sleep'. The same meaning in Hinukh could theoretically be explained as th result of secondary influence from East Tsezic, but actually all other West Tsezic verbs for 'to lie' and 'to sleep' can be analyzed as later introductions in individual lects. Thus, the easiest solution is to reconstruct Proto-Tsezic *class=itʼʷ 'to lie / to sleep', which was retained in East Tsezic and Hinukh, having been superseded by various (sometimes obscure) verbs in the majority of West Tsezic lects.

    It should be noted that Khwarshi ƛVs 'to sleep' is of unclear origin. The idea that it could be a compound of two verbs *ƛV- 'to sleep' [NCED: 619] and *-Vs 'to sleep' [NCED: 1037] (both are not attested elsewhere in Tsezic, although both possess appropriate North Caucasian comparanda) does not look apt.

    Replacements: {'to fall' > 'to lie'} (Kidero Dido); {'to be in the lateral position' > 'to lie'} (Sagada Dido).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary verbal stem with polysemy: 'to lie / to sleep'.

    Distinct from bɑcʼǝ {ба̇цIǝ} 'spleen' [Isakov & Khalilov 2001: 34; Kibrik & Kodzasov 1990: 35; van den Berg 1995: 287], erroneously glossed as 'liver' in [Bokarev 1961: 151, 178].

    Distinct from bocʼe {боцIе} 'spleen' [Khalilov & Isakov 2005: 102; Kibrik & Kodzasov 1990: 35].

    Kwantlada Khwarshi: žubu {жубу} 'liver' [Karimova 2014; Khalilova 2009: 6].

    Reconstruction shape: Correspondences seem regular except for the vowel metathesis in e-u > u-e in Proto-West Tsezic.

    Semantics and structure: Primary substantive root.

    Kwantlada Khwarshi: class=uχala {лухала} 'long' [Karimova 2014; Khalilova 2009: 99, 102, 251]. Polysemy: 'long (spatial) / long (temporal) / tall (of person)'.

    Reconstruction shape: Correspondences seem regular except for some vowel peculiarities.

    Semantics and structure: Primary adjective stem with polysemy: 'long (spatial) / long (temporal) / tall (of person)'.

    Distinct from käkä {каькаь}, enigmatically glossed as 'small louse' in [Khalilov 1995: 144] (a nursery word?).

    Distinct from kaki {каки}, enigmatically glossed as 'small louse' in [Khalilov & Isakov 2005: 192] (a nursery word?).

    Distinct from ħani {хIани}, enigmatically glossed as 'small louse' in [Khalilov 1999: 263].

    Kwantlada Khwarshi: noco {ноцо} 'louse' [Karimova 2014].

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Primary substantive root.

    Distinct from the adjective buχ-d-u ~ buχ-tʼ-u 'male (n., adj.)' [Isakov & Khalilov 2001: 46; Kibrik & Kodzasov 1990: 220; van den Berg 1995: 289].

    Distinct from betʼerhan {бетIергьан} with polysemy: 'master / husband / Lord' [Isakov & Khalilov 2001: 37; Kibrik & Kodzasov 1990: 59; van den Berg 1995: 287], borrowed from Avar betʼˈerhan 'husband / Lord'.

    A second candidate is the loanword bikinči {бикинчи} with polysemy: 'man / brave man, fine young man' [Khalilov 1995: 46, 317]. This word seems more marginal in the meaning 'man' than the inherited form abo, because in almost all the attested examples for 'man', Bezhta uses abo: “Men have come to see us” [Khalilov 1995: 24], “brave man” [Khalilov 1995: 83], “The men talked until the midnight” [Khalilov 1995: 107], “Quarrel is not man's matter” [Khalilov 1995: 124], “The men were walking around” [Khalilov 1995: 172].

    Distinct from borrowed bikinab {бикинаб} 'male' [Khalilov 1995: 46] and betʼerhan {бетIергьан} with polysemy: 'master / husband / Lord' [Khalilov 1995: 44; Madieva 1965: 150].

    Distinct from the borrowing bikinab 'male' [Kibrik & Kodzasov 1990: 220] and betʼerhän 'husband' [Kibrik & Kodzasov 1990: 59].

    Distinct from the inherited form abo 'father' [Kibrik & Kodzasov 1990: 54].

    Distinct from the borrowing bikinäb 'male' [Kibrik & Kodzasov 1990: 220] and betʼerhan 'husband' [Kibrik & Kodzasov 1990: 59].

    Distinct from the inherited form abo 'father' [Kibrik & Kodzasov 1990: 54].

    Cf. also the borrowing bikinči {бикинчи} with polysemy: 'man / brave man' [Khalilov & Isakov 2005: 82] (< Avar bixinči 'man').

    Distinct from the borrowed term bikinaw 'male / brave' [Khalilov & Isakov 2005: 82; Kibrik & Kodzasov 1990: 220] < Avar bixina-b 'male'.

    Distinct from two terms for 'husband': inherited χodːo {ходдо} [Khalilov & Isakov 2005: 354; Kibrik & Kodzasov 1990: 59], and borrowed betʼerhan ~ betʼerħan {бетIергьан, бетIерхIан} 'husband / brother-in-law (husband's brother) / Lord' [Khalilov & Isakov 2005: 73] (< Avar betʼerhan 'husband / Lord').

    There is also a borrowed term: bikinči {бикинчи} with polysemy: 'man / brave man' [Khalilov 1999: 56] (< Avar bixinči 'man').

    Distinct from the inherited term gulu-či 'male' [Khalilov 1999: 86; Kibrik & Kodzasov 1990: 220] and the borrowed term bikinaw 'male / brave' [Khalilov 1999: 56] < Avar bixina-b 'male'.

    Distinct from two terms for 'husband': inherited χediyu ~ χedyu {хедийу} [Khalilov 1999: 246; Kibrik & Kodzasov 1990: 59] and borrowed betʼerhan {бетIергьан} 'husband / master / Lord' [Khalilov 1999: 48] (< Avar betʼerhan 'husband / Lord').

    Distinct from χol 'husband' [Sharafutdinova & Levina 1961: 95].

    Distinct from Inkhokwari χol 'husband' [Kibrik & Kodzasov 1990: 59]. It should be noted that in the only instance in [Bokarev 1959: 161], χol is used in the meaning 'man'.

    Kwantlada Khwarshi: žikʼo {жикIо} 'man' [Karimova 2014; Khalilova 2009: 8, 14, 42].

    Distinct from Kwantlada χol 'husband' [Khalilova 2009: 15].

    In Bezhta proper, retained in the meaning 'person', but superseded by the inherited *ʔɔbu A 'father' [NCED: 217] in the meaning 'man'.

    Replacements: {'father' > 'man'} (Bezhta proper).

    Reconstruction shape: Although all the forms are obviously related, reflexes of the initial consonant are quite irregular: *z- in East Tsezic, ž- in West Tsezic, besides that, r- in Hinukh and h- in Inkhokwari Khwarshi (in the latter case, the laryngeal onset can be explained by Andian influence).

    Semantics and structure: Perhaps a primary substantive root with polysemy: 'man / person'.

    Differently in [NCED: 336, 579], where it is analyzed as a compound (*ž-ikʼʷǝ ~ *z-ikʼʷǝ), which consists of two words for 'man' (both with North Caucasian etymology). The root *ž- ~ *z- is not attested elsewhere in Tsezic, but according to [NCED: 579], Hinukh rekʼʷe and Inkhokwari Khwarshi hikʼo only reflect the second part of the compound < Proto-Tsezic *rikʼʷǝ ~ *hikʼʷǝ (< Pre-Proto-Tsezic *hirkʼʷV). However, such a scenario faces serious difficulties: besides strange phonetic development, it is unclear how plain *hikʼʷǝ could survive in Inkhokwari Khwarshi, when other Khwarshi dialects have *žikʼʷǝ; the same objection is applicable to *rikʼʷǝ in Hinukh.

    There are two synonymous adverbs in Hunzib, both with polysemy: 'many / much / often' - telːi and laχːi.

    Cf. the discovered examples for telːi 'many': “many errors” [Isakov & Khalilov 2001: 58], “Many goods were delivered in the store” [Isakov & Khalilov 2001: 153], “there were many people from Tladal” [van den Berg 1995: 180].

    Cf. the discovered examples for laχːi 'many': “there are many goats on our mountain” [Isakov & Khalilov 2001: 27], “there are many people” [Isakov & Khalilov 2001: 56], “We heard many songs” [Isakov & Khalilov 2001: 134], “They took many things in the warehouse” [Isakov & Khalilov 2001: 149].

    We treat telːi and laχːi as synonyms.

    A third candidate is class=ižeqʼ with the same polysemy: 'much / many / often' [van den Berg 1995: 307; Bokarev 1961: 156], but this one seems more marginal. The only discovered example is “We don't often (r=ižeqʼ) eat beet” [Isakov & Khalilov 2001: 178]. Historically, =iže-qʼ although the suffixal pattern is not entirely clear.

    There are multiple Bezhta equivalents for the meanings 'many' or 'much' listed in [Khalilov 1995: 316]. Browsing through the source demonstrates that teli is definitely the most frequent word for the meaning 'many'. Cf. some examples: “The woodcutter has chopped a lot of wood” [Khalilov 1995: 31], “There are many anxieties in the soul” [Khalilov 1995: 50], “People have a lot of cattle” [Khalilov 1995: 51], “There are many people in the railway station” [Khalilov 1995: 53], “a lot of clothes” [Khalilov 1995: 55], “a lot of smoke” [Khalilov 1995: 63], “to prepare many bundles of wood” [Khalilov 1995: 64], “There were many mistakes in the dictation” [Khalilov 1995: 66], “many pebbles” [Khalilov 1995: 67], “many children”, “Many years have passed” [Khalilov 1995: 237], and so on.

    Out of the other expressions for 'many', one should mention:

    1) äƛä-š {аьлIаьш} with polysemy: 'many / rural, of village' [Khalilov 1995: 32]; the examples: “many people”, “many books”; derived from äƛ 'village, aul; many people, crowd', i.e. 'many' as 'crowd-like';

    2) öl-lö {оьллоь} [Khalilov 1995: 212]; the examples: “I have a lot of books (lit.: There are a lot of books of mine)”, “There is a lot of earth tilled with tractor”; regular past tense from öl 'to be sufficient' [Khalilov 1995: 211].

    3) ladi {лади} [Khalilov 1995: 316; Madieva 1965: 173], missing from the main section of [Khalilov 1995].

    Out of a substantial number of more specific or marginal Hinukh terms for 'many' and 'much' (both inherited and borrowed), one should also mention raɬad {ралъад} with polysemy: 'sea / many' [Khalilov & Isakov 2005: 299] (< Avar raɬad 'sea').

    Distinct from more marginal or specific terms:

    1) ʡaƛo-s {гIалIос} with polysemy: 'many / rural, of village' [Khalilov 1999: 105], derived from ʡaƛ 'village, aul; many people, crowd', i.e. 'many' as 'crowd-like';

    2) χomi {хоми} 'much' [Khalilov 1999: 249];

    3) ʡuraw {гIурав} with polysemy: 'enough / many' [Khalilov 1999: 111], borrowed from Avar dial. ʡura-w 'ripe'.

    Two different Sagada words for 'many' are offered in [Abdulaev 2014]: ʡaši in the context "There are many stones on the ground" and class=aqʼˤu in the context "I have many friends". We have to treat both forms as synonyms, although it should be noted that Abdullaev's ʡaši in the context "There are many stones on the ground" could actually express a concept of a thick layer of stones.

    A second, apparently more marginal, expression for 'many' is dacːa-n {даццан} [Karimova 2014], which is quoted as daca-n 'completely, in full' in [Sharafutdinova & Levina 1961: 121].

    A second, apparently more marginal Inkhokwari expression for 'many' is doco-n {доцон} [Karimova 2014].

    Kwantlada Khwarshi: ʕezaʕan ~ ʕezeʕan ~ eˤzaʔan {гIезагIан, эIзааIн} 'many' [Karimova 2014; Khalilova 2009: 132]. Polysemy: 'many / much'. Cf. the examples: "He did many things for his friend to stay" [Khalilova 2009: 418], "The boy became happy to get many presents" [Khalilova 2009: 468].

    A second, apparently more marginal, Kwantlada expression for 'many' is doco-n {доцон} [Karimova 2014; Khalilova 2009: 283]. Cf. the example: "many mullahs from villages" [Khalilova 2009: 283].

    In East Tsezic (Hunzib, Bezhta), 'many' is expressed by the form *teli. Formally, *teli can be reconstructed as the Proto-East Tsezic term for this meaning. It must be noted, however, that East Tsezic *teli lacks any etymology and therefore represents a potential loanword. In Tlyadal Bezhta, *teli was superseded by class=ä̃tʼä 'many', which looks inherited, but seems to lack any Tsezic etymology.

    In Hinukh and the Dido dialects two words with the meaning 'many' occur in a "criss-crossed" interdialectal situation: class=aqʼˤu and (ʡ)aši. In both cases, 'many' is a secondary meaning: class=aqʼˤu originates from Proto-Tsezic *class=uqʼˤV 'big' q.v., (ʡ)aši < Proto-Tsezic *ʕašː- 'thick (2D)' [NCED: 608] (*ʕašː- retaines the meaning 'thick', e.g., in Kidero Dido).

    Topologically, the best candidate for the Proto-Tsezic status is *class=iˤžV ~ *-žː- B [NCED: 653], which is attested as the plain stem 'big' in Hinukh and Dido and as the suffixed stem 'many' in Khwarshi (*eˤz-ya-ʔan) and Hunzib (class=iže-qʼ). But the fact that Khwarshi and Hunzib demonstrate different suffixal patterns suggests that we deal with late independent introductions in Khwarshi and Hunzib. The original meaning of *class=iˤžV remains unclear (Hinukh & Dido 'big' q.v. is likewise secondary).

    From the external point of view, the best candidate for the Proto-Tsezic status is *laχːi (~ -ʁ-), attested as isolated Hunzib laχːi 'many', but having possible North Caucasian comparanda with the meaning 'many'. However, it is hard to suppose that *laχːi survived with its original meaning 'many' only in Hunzib.

    Formally, we fill the Proto-Tsezic slot with East Tsezic *teli, which lacks North Caucasian comparanda.

    Replacements: {'village, crowd' > 'many'} (Bezhta proper, Kidero Dido); {'to be sufficient' > 'many'} (Bezhta proper); {'sea' > 'many'} (Hinukh); {'thick (2D)' > 'many'} (Hinukh, Sagada Dido); {'big' > 'many'} (Kidero Dido, Sagada Dido); {'how many?' > 'many'} (Khwarshi proper, Inkhokwari Khwarshi, Kwantlada Khwarshi).

    Reconstruction shape: Correspondences seem regular.

    Semantics and structure: Probably an adverb form.