Build tree| Use tables | |||||||||||||
| Multiple windows | |||||||||||||
| Morphology/dictionary lookup window | |||||||||||||
| Show languages in map | |||||||||||||
| Encoding |
| ||||||||||||
Build treeCaucasian Albanian: Unattested.
Tsakhur-Kum Tsakhur: kuma [Kibrik & Kodzasov 1990: 207].
The same in the Kurag subdialect: kum 'smoke' [Magometov 1970: 206 sentence 9].
The same in the Khanag subdialect: kum 'smoke' [Uslar 1979: 787, 993; Dirr 1905: 185, 228]. The same in the Khyuryuk subdialect: kum {кум} 'smoke' [Genko 2005: 89].
The same in the Khiv subdialect: kum {кум} 'smoke' [Genko 2005: 89]. The same in Literary Tabasaran: kum {кум} 'smoke' [Khanmagomedov & Shalbuzov 2001: 203].
The same in Literary Lezgi: gum {гум} 'smoke' [Talibov & Gadzhiev 1966: 88; Gadzhiev 1950: 191; Haspelmath 1993: 489, 526].
The same in the Akhty dialect: Khlyut gɨm 'smoke' [Kibrik & Kodzasov 1990: 207].
In Archi, the meaning 'smoke' is expressed with *ʔʷɨqʼː [NCED: 251], an isolated form within Lezgian, but with external North Caucasian comparanda in the meaning 'fumes, stink'.
In Proto-Nuclear Lezgian, *kːunː was superseded with *ƛuma (~ *ƛʷi-) [NCED: 590], which is retained with the meaning 'smoke' in all Nuclear Lezgian lects. Its original Proto-Lezgian meaning cannot be established, because it was lost in Udi and Archi, but external North Caucasian comparison suggests semantics of 'wind' or 'air'.
Replacements: {'smoke' > 'dust'} (Archi).
Reconstruction shape: Correspondences seem regular except for the Udi diphthong.
Semantics and structure: Primary substantive root.
Distinct from ayz-esun {айзесун} 'to stand up' [Gukasyan 1974: 38; Fähnrich 1999: 6].
Caucasian Albanian: bur-esun with polysemy 'to stand (of humans, things) / to be at hand / to be, remain / to remain, stay / to dwell' [Gippert et al. 2008: IV-12]. As plausibly analyzed in [Gippert et al. 2008: II-45, IV-12], this is a secondary verb based on the nominal form buri 'having come to be' (< *bu-ar-i), a stative past participle from the generic verb bu- 'to be' [Gippert et al. 2008: IV-11].
Distinct from the nursery word ʁˈetʼ-bo- 'to stand' [Kibrik et al. 1977b: 311, 383; Chumakina et al. 2007] (complex verb, formed with the suppletive light verb -bo- 'to say').
Distinct from =χːˈa- 'to stand up; to grow; to be in heat (of animals)' [Kibrik & Kodzasov 1988: 85; Kibrik et al. 1977b: 337, 354] and yˈatːi =χːa- 'to stand up' [Kibrik et al. 1977b: 254, 354] (yˈatːi 'up').
Distinct from qː=uzur- 'to stand up' [Kibrik & Kodzasov 1988: 85].
In [Comrie & Khalilov 2010: 561], 'to stand' is incorrectly glossed as the analytic form aʁmi yixar {агъми йихьар}, which, in fact, means 'to stand face to face, to say smth. right to smb.'s face' [Meylanova 1984: 16].
Distinct from qː=uzar- [imperf.] / qː=uzur- [perf.] 'to stand up' [Kibrik & Kodzasov 1988: 85; Meylanova 1984: 91, 208].
Distinct from Mikik ozaq-Vx- 'to stand up' [Kibrik & Kodzasov 1988: 85] (the second element could be the auxiliary verb eːxe- [imperf.] / ɨxa- [perf.] 'to become' [Kibrik & Kodzasov 1988: 62]).
Initial Vlʸ= is a prefix with general semantics [Ibragimov 1990: 123; Kibrik & Kodzasov 1988: 41].
Initial l= and χˤɨ=ʁ= are prefixes with general semantics [Ibragimov 1978: 95; Alekseev 1994a: 227; Makhmudova 2001: 165].
The same in the Khudig subdialect: a=H=z-a- 'to stand' [Shaumyan 1941: 191].
The same in the Usug subdialect: ʁ=uz-a- 'to stand' [Shaumyan 1941: 191].
Note a rare case of retention of the Lezgian imperfective infix -r- in the Gequn and Proper Aghul imperfective stem ʁ=u-r-z-a- (cf. [Suleymanov 1993: 138 f.]).
The same in the Kumi subdialect: di=yiqː-ˈ {дийикъув} 'to stand' [Genko 2005: 60].
A different pattern in the Khanag subdialect: di=ʁ=iǯʷ-ˈ [imperf.] / du=ʁ=uǯʷ-ˈ [perf., inf.] with polysemy: 'to stand / to stand up / to stay at rest' [Uslar 1979: 673, 1007; Dirr 1905: 164, 243]. The change u > i in the imperfective stem is synchronically regular, see [Kibrik & Kodzasov 1988: 34 f.] for the same ablaut in Dyubek.
The same in the Khyuryuk subdialect: di=ʁ=iǯʷ-ˈ [imperf.] / du=ʁ=uǯʷ-ˈ [perf.] {дугъубжъув} with polysemy: 'to stand / to stay at rest' [Genko 2005: 61].
The same in Literary Tabasaran: di=yiʁ-ˈ {дийибгъуб} 'to stand; to stop moving' [Khanmagomedov & Shalbuzov 2001: 147].
Differently in the Khiv subdialects: ʁu=d=uǯʷ-ˈ {гъудужъуб} 'to stand; to stand up' [Genko 2005: 44]. Cf. Khiv du=ʁ=uǯʷ-ˈ {дугъужъуб} 'to stop moving' [Genko 2005: 61].
Formally, it is possible to reconstruct the Proto-Tabasaran opposition =ʔiɢ- (=yiʁ-) 'to stand' / =iǯʷ- (=uǯʷ-) 'to stand up' (retained in Dyubek and Kondik, but secondarily lost in favor of the latter root in most of the other dialects). External Lezgian comparison suggests, however, that =iǯʷ- (=uǯʷ-) should be rather reconstructed with Proto-Tabasaran polysemy 'to stand / to stand up'; if so, Dyubek and Kondik =ʔiɢ- (=yiʁ-) 'to stand' is a late introduction.
The same in Literary Lezgi: aqːʷ=ˈaz- {акъвазун} with polysemy: 'to stand (applied to the animated subject) / to stop moving / to stay at rest' [Talibov & Gadzhiev 1966: 38; Gadzhiev 1950: 822; Haspelmath 1993: 481, 527; Gyulmagomedov 2004, 1: 58]. Distinct from literary qː=arˈaʁ- {къарагъун} 'to stand up' [Talibov & Gadzhiev 1966: 180; Gadzhiev 1950: 111; Haspelmath 1993: 501; Gyulmagomedov 2004, 1: 394].
For the phonetics cf. Migrakh (subdialect of the Doquzpara dialect < Samur group) q=ucː(ˈV)- 'to stand' [Meylanova 1964: 258], Khuryug (subdialect of the Akhty dialect < Samur group) q=ucː(V)- 'to stand' [Meylanova 1964: 315].
Initial aqːʷ=, qː=, q= are old spatial prefixes.
Nevertheless, the Proto-Lezgian root *ʔecːʷär- [NCED: 1025] can be assuredly reconstructed with polysemy: 'to stand / to stand up'. Actually, the isogloss of lexical discrimination between the two meanings seems to be a recent areal innovation among Lezgian languages, because newly introduced verbs with the specific meaning 'to stand' or 'to stand up' almost never coincide between languages.
The root *ʔecːʷär- [NCED: 1025] is retained with polysemy 'to stand / to stand up' in some Nuclear Lezgian languages: West Lezgian (Tsakhur, Rutul), Aghul and perhaps in Proto-Tabasaran (although the latter is not certain). In one of the outliers - Archi - as well as in Lezgi, the original meaning was narrowed to 'to stand'. On the contrary, *ʔecːʷär- was narrowed to 'to stand up' in the second outlier - Udi - and in South Lezgian (Kryts, Budukh). Similarly, the best formal solution for Proto-Tabasaran would be to reconstruct *ʔecːʷär- with the narrow meaning 'to stand up'.
The new verbs for 'to stand' are:
1) *ʔVtʼVl-, attested in South Lezgian (Kryts, Budukh) as 'to stand', lacking further etymology;
2) in Udi, the etymologically obscure root čur- is used for the verb 'to stand' (formally the same root as in the Udi verb 'to walk');
3) in Caucasian Albanian, 'to stand' is derived from the generic verb 'to be'.
The new verbs for 'to stand up' are:
1) *ʔiχːa- [NCED: 575]; this root means 'to stand up' in Luchek Rutul (theoretically, such a meaning could be reconstructed for Proto-Rutul) and in one of the outliers - Archi. The root *ʔiχːa- was lost in the rest of Lezgian lects, so its Proto-Lezgian meaning is not reconstructible. Formally, the match between Archi and Rutul could yield the Proto-Lezgian root for 'to stand up', but we consider the available data too scant for such a reconstruction. Apparently Archi and Luchek Rutul represent independent innovations;
2) *yaqːV- [NCED: 275], attested as 'to stand up' in some Tabasaran dialects (formally this should be reconstructed as the Proto-Tabasaran verb 'to stand up'); the original Lezgian meaning of *yaqːV- was '(to be) high';
3) Etymologically, not entirely clear are Mikik Tsakhur ozaq-Vx- 'to stand up' and Lezgi qː=arˈaʁ- 'to stand up'.
Replacements: {'to be' > 'to stand'} (Caucasian Albanian), {'to be high' > 'to stand'} (Tabasaran).
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Primary verbal root with polysemy 'to stand / to stand up'.
In [Schulze 2001: 299] it is tentatively proposed to derive mučːˤuli from the Udi substantive mi 'cold, frost' ('star' as 'cold (light)'), but this is problematic from the morphologic point of view and seems impossible typologically.
In [Schulze 2001: 309] qabun is ad hoc segmented as qa-b-un and connected to the basic Lezgian term for 'star': *χːˤanːa [LEDb], which is impossible phonetically.
Caucasian Albanian: χalʸum [Gippert et al. 2008: IV-22]. The final -m looks like a fossilized plural exponent. This term probably continues Proto-Lezgian *χːˤanːa 'star'. It must be noted that the l-like value of sign #22 (currently transliterated as lʸ) is established with relative assurance, see [Gippert et al. 2008: II-13]. The development of Lezgian *nː > Caucasian Albanian lʸ is somewhat surprising (note that Lezgian *nː > Udi n), although not impossible.
Tsakhur-Kum Tsakhur: χˤanʸe [Kibrik & Kodzasov 1990: 198].
The same in the other subdialects: Arsug, Khudig Had 'star' [Suleymanov 1993: 69].
The same in the other subdialects: Tsirkhe χˤar, Duldug, Khpyuk Had 'star' [Suleymanov 1993: 57, 69; Shaumyan 1941: 180].
The same in the Khanag subdialect: χːˤar 'star' [Uslar 1979: 956, 994; Dirr 1905: 216, 229]. The same in other subdialects: Khyuryuk χːˤar {ххяр}, Chuvek χaǯ {хажж} 'star' [Genko 2005: 162, 168].
The same in other subdialect: Khiv χˤaǯ {хежж, хяжж}, Turag χˤad {хяд} 'star' [Genko 2005: 164, 168]. The same in Literary Tabasaran: χˤad {хяд} 'star' [Khanmagomedov & Shalbuzov 2001: 319].
Differently in two subdialects of the Eteg cluster: Tinit, Dzhikhtig čʼiz {чIиз} 'star' [Genko 2005: 186].
The same in Literary Lezgi: ʁed [abs.] / ʁetː-rˈe- [obl.] / ʁetː-ˈer [pl.] {гъед} 'star' [Talibov & Gadzhiev 1966: 93; Gadzhiev 1950: 243; Haspelmath 1993: 490, 527].
The same in the Akhty dialect: Khlyut χäd [abs.] / χäd-rˈa- [obl.] / χätː-ˈar [pl.] 'star' [Kibrik & Kodzasov 1990: 198].
In Udi, two etymologically obscure forms for 'star' occur: mučːˤuli, qabun. In Southern Tabasaran (Tinit, Dzhikhtig), 'star' is denoted with čʼiz, whose origin is likewise unclear.
The reconstruction of the Proto-Lezgian *χːˤanːa ~ *χːˤanː seems unproblematic, but the fact that *χːˤanːa ~ *χːˤanː lacks any external North Caucasian cognates is suspicious.
Actually, Proto-Lezgian *χːˤanːa ~ *χːˤanː looks like a derivative stem from the Proto-Lezgian substantive *χːˤanː 'fish' q.v. [NCED: 1078]. Cf. the data from two dialects, where both Proto-Lezgian terms survived: Shinaz Rutul χˤat (< *χˤad), Gyune Lezgi ʁed 'fish' vs. Common Rutul χˤad-ey, Gyune Lezgi ʁed 'star'.
Theoretically, it is possible to hypothesize a Proto-Lezgian mythologem, according to which stars are considered "sky fishes". The modern Archi complex expression for 'star' might be a clue to such a reconstruction: χːˤolˈoš-qˤan, literally 'qˤan from above / from the sky'. In modern Lezgian lects, 'fish' is a very unstable word (as opposed to 'star'); this could be explained as the result of later attempts to avoid the homonymy 'star' / 'fish' already after the myth of "sky fishes" had disappeared. If so, the data from Eteg Tabasaran are particularly interesting, because Eteg is the third dialect among Lezgian languages which retains the old term for 'fish' (Tinit, Dzhikhtig χˤad), but it is exactly in Eteg where the old term for 'star' has been superseded with the unclear word čʼiz.
Replacements: {'fish' > 'star'(?)} (Proto-Lezgian).
Reconstruction shape: Correspondences seem regular, except for dissimilative qˤ (for expected χːˤ) in Archi and ejectivization of the final consonant in South Lezgian (Kryts, Budukh); note also the somewhat strange reflex *nː > lʸ in Caucasian Albanian. The direct stem with the final vowel *χːˤanːa (a rare nominal type) is reconstructed on the basis of the Tsakhur form χˤanʸe [NCED: 171]; on the other hand, Tsakhur may represent the same suffixal formation as Rutul χˤad-ey (with the loss of -y in Tsakhur).
Semantics and structure: Nominal stem, maybe of denominative nature. The oblique stem is *χːˤanːä-.
Caucasian Albanian: žˤe [abs.] / žˤen- [obl.], see [Gippert et al. 2008: II-14, 23, IV-18].
Distinct from qʷan with polysemy 'large flat stone / large frying pan' [Kibrik et al. 1977b: 298; Chumakina et al. 2007].
Distinct from a more specific term dahar 'big stone' [Kibrik & Kodzasov 1990: 201] (ultimately of Persian origin, see notes on Alyk Kryts).
Distinct from inherited Alyk χud 'rock, cliff' [Authier 2009: 25, 39, 81].
Tsakhur-Kum Tsakhur: ɢaye [Kibrik & Kodzasov 1990: 201].
A second candidate is qːat {къат}, quoted in [Ibragimov 1978: 118] as a synonym for duχul, but not observed in other sources.
A third candidate is cʼʷar, glossed as 'stone, small stone, gravel' in [Dirr 1912: 37, 179, 192], but, judging by Dirr's examples, this word seems more rare and marginal.
Cf. also non-inherited dahar '(building) stone' [Ibragimov 1978: 27, 29, 118].
A second candidate is dahˤar {даIгьаIр} 'stone' [Dzhamalov & Semedov 2006: 95, 346] (with the only example: "large stones"), ultimately borrowed from Persian (see notes on Alyk Kryts). The difference between duχul and dahˤar is unclear; it is possible that dahˤar specifically denotes a large stone.
In [Comrie & Khalilov 2010: 50], several specific terms are incorrectly quoted as synonyms for generic 'stone'.
Distinct from inherited duχul 'small stone' [Kibrik & Kodzasov 1990: 201] and qːat 'pebble' [Kibrik & Kodzasov 1990: 202].
The distribution suggests that duχul is the Proto-Rutul term for 'stone' (historically duχ-ul with the archaic suffix -l).
The same in the Arsug subdialect: Hʷan 'stone' [Suleymanov 1993: 54].
The same in the Khanag subdialect: qːan 'stone (in general)' [Uslar 1979: 818, 995]; note, however, that in [Dirr 1905: 186, 230] Khanag qːan is quoted only in the expression niqʼ-r-in qːan 'tombstone' (niqʼ 'tomb, grave'). According to [Dirr 1905: 161, 230], the generic Khanag term for 'stone' is ʁarʒ, which is glossed as 'rock, cliff' in [Uslar 1979: 651] - perhaps a natural semantic rebuilding during the 2nd half of the 19th century between Uslar's and Dirr' records.
The same in the Khyuryuk subdialect: qːan {къан} 'stone' [Genko 2005: 99]; distinct from Khyuryuk ʁarʒ {гъарзз} 'rock, cliff' [Genko 2005: 41].
The same in the Tinit subdialect: qːʷan {къоан} 'stone' [Genko 2005: 101] (phonetically rather a Northern Tabasaran form!).
The same in Literary Tabasaran: ʁʷan {гъван} 'stone' [Khanmagomedov & Shalbuzov 2001: 110]; distinct from literary ʁarz {гъарз} 'rock, cliff' [Khanmagomedov & Shalbuzov 2001: 108].
Two terms for 'stone' are documented for the Khiv subdialect: ʁarz {гъарз} with polysemy: 'stone / rock, cliff / stony slope / precipice' [Genko 2005: 41] and ʁʷan {гъоан} 'stone' [Genko 2005: 43]. The difference is unknown.
The same in Literary Lezgi: qːʷan {къван} 'stone' [Talibov & Gadzhiev 1966: 184; Gadzhiev 1950: 278; Haspelmath 1993: 503, 527].
The same in the Akhty dialect: Khlyut qːʷan 'stone (in general)' [Kibrik & Kodzasov 1990: 201]. Distinct from the Khlyut more specific term kertːʷeχ 'formless stone' [Kibrik & Kodzasov 1990: 201].
1) *qːʷan [NCED: 490]. This is the generic term for 'stone' in East Lezgian (Aghul, Tabasaran, Lezgi), but denotes 'large flat stone; large frying pan' in Archi (qʷan [Kibrik et al. 1977b: 298; Chumakina et al. 2007]). External North Caucasian comparanda point to the meanings 'flat stone' or 'large stone';
2) *χutː (~ *χː-) [NCED: 428]. This root denotes 'stone' in Kryts proper and Rutul (Rutul suffixed and metathesized duχ-ul), but 'rock, cliff' in Alyk Kryts. In [NCED: 428], Kryts proper χud-il 'tomb-stone' is also quoted (directly corresponds to Rutul duχ-ul 'stone'), but this Kryts form is not attested in [Kibrik & Kodzasov 1990: 191]. The root is not attested in the rest of Lezgian. Its external North Caucasian comparanda (if correct) point to the meaning 'grave, tomb';
3-4) in both of the outliers, etymologically obscure forms are attested: Udi-Caucasian Albanian žˤe (which implies Proto-Lezgian *č:en ~ -ä- ~ -nː) and Archi čʼelˈe.
Provisionally, we fill the Proto-Lezgian slot with *qːʷan.
It must be noted that in Tabasaran dialects, *qːʷan tends to be superseded with the form ʁarz in the meaning 'stone'; the Proto-Tabasaran meaning of ʁarz apparently was 'rock, cliff'.
In Alyk Kryts and Luchek Rutul, the Persian loanword dahar is attested with the late semantic development 'big stone, rock, cliff' > 'stone (in general)'.
In Budukh and Tsakhur, inherited terms were superseded with Azerbaijani loanwords.
Replacements: {'big stone, rock, cliff' > 'stone'} (Alyk Kryts, Luchek Rutul), {'rock, cliff' > 'stone'} (Tabasaran dialects).
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Primary substantive root. The oblique stem is not reconstructible.
Caucasian Albanian: bʕeʁ [Gippert et al. 2008: IV-11].
Tsakhur-Kum Tsakhur: wireʁ [Kibrik & Kodzasov 1990: 197].
The same in the Khanag subdialect: riqː 'sun' [Uslar 1979: 896, 1006] (in [Dirr 1905: 203, 242], quoted as reʁ 'sun' - either actually a form from some Southern Tabasaran subdialect or the beginning of the phonetic process qː > ʁ in Khanag during the 2nd half of the 19th century between Uslar's and Dirr's records).
The same in the Khyuryuk subdialect: riqː {рикъ} 'sun' [Genko 2005: 133].
The same in the Khiv subdialect: riʁ ~ reʁ {ригъ, регъ} 'sun' [Genko 2005: 132]. The same in Literary Tabasaran: riʁ {ригъ} 'sun' [Khanmagomedov & Shalbuzov 2001: 260].
The same in Literary Lezgi: raʁ [abs.] / raqː-ˈini- [obl.] {рагъ} 'sun' [Talibov & Gadzhiev 1966: 270; Gadzhiev 1950: 798; Haspelmath 1993: 504, 527].
The same in the Akhty dialect: Khlyut raʁ [abs.] / reqː-ˈina- [obl.] / raqː-ˈar [pl.] with polysemy: 'sun / sunny days (pl.)' [Kibrik & Kodzasov 1990: 197].
Replacements: {'sun' > 'sunny day (pl.)'} (Akhty Lezgi).
Reconstruction shape: Basic correspondences are regular, although the original bisyllabic structure tends to be simplified in many languages.
Semantics and structure: Primary substantive root. The oblique stem could be *wirɨqːV-.
Caucasian Albanian: Unattested.
In [Dirr 1908: 188, 217] 'to swim' is quoted as {хwа-к̠}, which should be interpreted as something like χʷakʼ ~ χːʷakʼ - apparently the same verb as χːʷˈaːkʼa- 'to walk around looking for something' [Kibrik et al. 1977b: 339; Kibrik et al. 1977a 3: 235; Chumakina et al. 2007].
For the Khyuryuk subdialect only the noun čʼučʼurufˈan {чIучIуруфан} 'swimming' is documented [Genko 2005: 187].
Differently in the Tinit subdialect: lepˈe yeb- {лепе ебуб} 'to swim', literally 'wave' + 'to hit, beat' [Genko 2005: 114].
There also exists a more complex expression: lepˈe kːada-ˈuri ʁˤaʁ- {лепе ккадаури гъягъюб} 'to swim' [Genko 2005: 114], literally 'to go (ʁˤaʁ-), clearing away (kːada-) the wave (lepˈe)'; Genko's note is not quite certain, but apparently this is a Khiv form.
Two expressions are found in Literary Tabasaran: sirnˈaw apʼ- {сирнав апIуб} 'to swim', literally 'swimming' + 'to do' [Khanmagomedov & Shalbuzov 2001: 272]; lepˈe yiw- {лепе йивуб} 'to swim', literally 'wave' + 'to hit, beat' [Khanmagomedov & Shalbuzov 2001: 217].
The same in Literary Lezgi: sirnˈaw awˈu- or compressed sirnˈaw- {сирнав авун, сирнавун} 'to swim' [Talibov & Gadzhiev 1966: 294; Gadzhiev 1950: 539; Haspelmath 1993: 527].
1) Archi analytic ɬːan-a- 'to do water';
2) Rutul analytic xed haʔ- 'to do water', an exact parallel to the complex verb in Archi;
3) the verb 'to walk around looking for something' in archaic Archi;
4) Northern Tabasaran čʼučʼurufan apʼ- 'to do čʼučʼurufan'; perhaps onomatopoeic.
We follow the formal Archi-Nuclear Lezgian (Rutul) match and reconstruct the virtual idiom *ɬːänː ʔaʔa(r)- 'to do (ʔaʔar-) water (*ɬːänː)' [NCED: 257, 1060] as the Proto-Lezgian expression for 'to swim', although it is very probable that Archi and Rutul formations represent late and independent introductions.
Replacements: {'to do water' > 'to swim'} (Archi, Rutul), {'to beat a wave' > 'to swim'} (Southern Tabasaran).
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Analytic expression noun + auxiliary verb.
Distinct from tːuntːuz 'bird's tail, rump' [Fähnrich 1999: 32].
Caucasian Albanian: Unattested.
There exists another word for 'tail': qačʼ, quoted only once in [Authier 2009: 39]. According to G. Authier’s pers. com., this is a rare specific term, which denotes 'long, lean tail containing bones'. Alyk qačʼ resembles the basic Khinalugh term qʼaž [abs.] / qʼačʼ- [obl.] 'tail' (note that, first of all, the normal direction of borrowing is Lezgian > Khinalugh).
Distinct from ǯɨkʼrɨ 'tail of cloven-hoofed animal, horse's tail' (sic?) [Ibragimov & Nurmamedov 2010: 161] and aˤrd 'fat tail of sheep (курдюк); fatty meat' [Ibragimov & Nurmamedov 2010: 62] (the latter is borrowed from Azerbaijani ard 'back part of body'). Ibragimov & Nurmamedov's gloss 'tail of cloven-hoofed animal, horse's tail' for ǯɨkʼrɨ looks suspicious, because this word means 'fat tail of sheep (курдюк)' (a very different meaning) in Mikik and Gelmets, whereas in Mishlesh 'horse's tail' is denoted by generic bɨˤt(ʸ), see examples in [Kibrik et al. 1999: 822].
Tsakhur-Kum Tsakhur: bɨˤt [Kibrik & Kodzasov 1990: 26]. Distinct from aˤrd 'fat tail of sheep (курдюк)' [Kibrik & Kodzasov 1990: 26] (borrowed from Azerbaijani ard 'back part of body').
Distinct from ǯɨkʼrɨ 'fat tail of sheep (курдюк); handle' [Kibrik & Kodzasov 1990: 26; Dirr 1913: 158, 228] and from aˤrd 'fat tail of sheep (курдюк)' [Kibrik & Kodzasov 1990: 26] (the latter is borrowed from Azerbaijani ard 'back part of body').
Distinct from ǯɨkʼrɨ 'fat tail of sheep (курдюк); handle' and aˤrd 'fat tail of sheep (курдюк)' [Kibrik & Kodzasov 1990: 26] (the latter is borrowed from Azerbaijani ard 'back part of body').
The same in the other subdialects: Duldug, Tsirkhe, Khpyuk ruǯ 'tail' [Suleymanov 1993: 63; Shaumyan 1941: 160].
The same in the Khanag subdialect: riǯʷ with polysemy: 'tail / fat tail of sheep (курдюк) / handle' [Uslar 1979: 896, 1009; Dirr 1905: 203, 246]. The same in the Khyuryuk subdialect: riǯʷ {рижъ} with polysemy: 'tail / handle' [Genko 2005: 133].
The same in the Khiv subdialect: ruǯʷ {ружъ} with polysemy: 'tail / fat tail of sheep (курдюк) / handle' [Genko 2005: 134]. The same in Literary Tabasaran: riǯʷ {рижв} with polysemy: 'tail / fat tail of sheep (курдюк) / handle / hanger-on, stooge' [Khanmagomedov & Shalbuzov 2001: 260].
The same loanword in Literary Lezgi: tːum {тум, ттум} with polysemy: 'tail / fat tail of sheep (курдюк) / handle / confederate' [Talibov & Gadzhiev 1966: 315; Gadzhiev 1950: 912; Haspelmath 1993: 508, 527].
The same in the Akhty dialect: Khlyut tːɨm 'tail' [Kibrik & Kodzasov 1990: 26]. Khlyut ɨm < um is a late process, cf., e.g., inherited gɨm 'smoke'.
In Tsakhur, 'tail' is expressed with the etymologically obscure form bɨˤtːʸ-.
In Lezgi, apparently an Iranian loanword is used for 'tail'.
Replacements: {'tail' > 'handle, haft, grip'} (Budukh, Tsakhur, Rutul, Tabasaran), {'tail' > 'toady, hanger-on'} (Budukh, Rutul, Tabasaran).
Reconstruction shape: Correspondences seem regular, except for some peculiarities: metathesis (*Hɨrčːʷ > *rɨčːʷ) in Aghul and Tabasaran, and elimination of the direct stem in favor of the oblique one with subsequent reduction of the first syllable (Kryts, Budukh, Rutul, Tsakhur).
Semantics and structure: Primary substantive root. The oblique stem is *Hɨrč:ʷɨ-.
Nidzh mo / ko / šo ~ tːe (both forms šo and tːa are used in the absolutive, whereas only tːe is used in other cases);
Vartashen me / ka / tːe.
On the other hand, it is calculated in [Schulze 2008: 255 ff.; Schulze 2005: 237 ff. (3.2.9.3)] that the medial demonstrative attributive ka is used significantly less frequently than proximal me and distal tːe both in Nidzh and Vartashen texts. Thereby it is possible to analyze this as a system with the basic binary opposition *me 'this' / *tːe 'that'.
The Nidzh attributive pronoun šo is probably secondary in the attributive function. The normal function of both Nidzh šo ~ šo-no and Vartashen še-no ~ šo-no is the non-attributive pronouns '(s)he, it, they' [Gukasyan 1974: 276, 277].
Caucasian Albanian: Demonstrative attributive pronouns in the Swadesh function are too poorly documented in the palimpsests (cf. [Schulze 2008: 303 ff.; Gippert et al. 2008: II-38]).
For the sake of lexicostatistics, the list in [Gippert et al. 2008: II-66] fills the slot 'this' with -me and the slot 'that' with -tʼe-, -še-. In actuality, -tʼe- is a suffixal morpheme, added to the oblique stem of the neuter anaphoric pronoun [Gippert et al. 2008: II-38] and, accordingly, to nominal referentialised forms [Gippert et al. 2008: II-29]; -š- is found only in the damaged adverbial form ešol-oqoc 'from this side' and in the adverb e-še '(t)here(?)' [Gippert et al. 2008: II-38, IV-15] (in fact, the interpretation of the latter form is also quite uncertain, cf. [Gippert et al. 2008: VII-90 fn. 70]). In turn, -me is apparently found only in the proximal adverb e-me 'here' [Gippert et al. 2008: II-38, IV-15].
As described in [Kibrik et al. 1977a 2: 124] and [Kibrik 1994: 319], the system of Archi demonstrative attributive pronouns is ternary (excluding the vertical oriented forms): ya-class 'this (near the speaker)' / yˈa-mu-class 'this (near the addressee)' / to-class 'that (far from the speaker and addressee)'. Since the basic lexicon per se is not only anthropocentric, but eventually egocentric, we prefer to assume that the basic system is binary: ya-class 'this' (near the speaker)' / to-class 'that (far from the speaker and addressee)'. It should be noted that, according to [Dirr 1908: 29], these two are statistically the most frequently used demonstrative pronouns.
According to reports in [Talibov 2007: 121 f.] and [Meylanova 1984: 192], the Budukh system of demonstrative attributive pronouns is apparently ternary: ulu ~ olu 'this (near the speaker)' / al-am ~ am 'this (near the addressee)' / ala 'that (far from the speaker and addressee)'. The medium member (al-am ~ am 'this near the addressee') seems marginal, and we prefer to exclude it from the list.
Kryts cognates suggest that Budukh stems ulu 'this' and ala 'that' are to be analyzed as compounds u-lu, a-la. In each case both morphemes represent meaningful elements. It should be noted, however, that in substantivized forms (i.e. modified by class-prefixes) the second morphemes can be omitted, thus u-class ~ u-lu-class 'this' / a-class ~ a-la-class 'that' [Talibov 2007: 122; Meylanova 1984: 192].
In [Alekseev 1994: 267 f.], the demonstrative pronouns are confused: ala is erroneously quoted for 'this', ulu - for 'that'.
According to [Kibrik et al. 1999; Ibragimov 1990] and [Ibragimov & Nurmamedov 2010], the Mishlesh and Literary Tsakhur system of demonstrative attributive pronouns is ternary: i-n 'this (near, here)' / ma-n 'that (there)' / še-n 'that (far away)' (or ha=y-n / ha=ma-n / hoː=še-n with the emphatic proclitics ha- 'here' and hoː 'there'). Browsing through texts in [Kibrik et al. 1999] suggests that ma-n (ha=ma-n) probably means 'this', but it is difficult to make a definite choice between i-n (ha=y-n) and ma-n (ha=ma-n), so we treat them as synonyms for 'this'.
ma-n also functions as a personal pronoun of the 3rd p.
Tsakhur-Kum Tsakhur: according to [Schulze 1997: 39], the system of demonstrative attributive pronouns is ternary: i-n 'this (near, here)' / ma-n 'that (there, visible)' / še-n 'that (far away, invisible)'. Semantic and pragmatic nuances are unknown, however.
ma-n also functions as a personal pronoun of the 3rd p.
It is not easy to understand all the nuances of the Mukhad system of demonstrative attributive pronouns, based on the aforementioned sources. Apparently the basic opposition is binary: mi 'this' / ti 'that'. Both pronouns can be additionally supplemented with the emphatic morpheme he-: he=mi / he=ti.
There also exists, however, a third pronoun ha {гьа} (with the variant ha-d, where the final element seems to be a fossilized class exponent), which is used as both the personal pronoun of the 3rd p. '(s)he, it, they' and the demonstrative attributive pronoun. The exact meaning of the demonstrative ha is not entirely clear. In [Alekseev 1994a: 225], ha is glossed as 'this, that'; in [Ibragimov 1978: 81], as 'that (distal deixis)'; on the contrary, in [Makhmudova 2001: 170] ha is explained as 'that (near the addressee)'. In any case, ha is, apparently, rarely used in the attributive meaning; according to [Dirr 1912: 38], the most frequent function of ha is the 3rd p. pronoun '(s)he, it, they'.
Dzhamalov & Semedov do not provide much information, but, apparently, the Ixrek system is the same as the Mukhad one: mi-dɨ 'this' / ti-n-dɨ 'that', plus ha-dɨ, which normally means '(s)he, it, they', but can also be used as the demonstrative attributive 'that'. The final -dɨ is the attributive suffix; -n- in ti-n-dɨ 'that' originates from the oblique stem.
Similarly in the Arsug subdialect: mi-m 'this' / ti-m 'that' / gi-m 'that (below)' / li-m 'that (above)' [Suleymanov 1993: 128, 130].
Similarly in the Khudig subdialect: me-d 'this' / ti-m 'that' / gi-m 'that (below)' / li-m 'that (above)' [Suleymanov 1993: 128, 130; Suleymanov 2003: 132].
It is reported in [Suleymanov 1993: 129] that in many dialects the pronouns 'that', 'that (below)' and 'that (above)' can be modified with the iconic proclitic oː= ~ hoː= ~ woː= (depending on the dialect), which expresses distal deixis, e.g., Keren hoː=ti 'that (far horizontally) / hoː=gi 'that (far below) / hoː=li 'that (far above). Apparently these forms are rare and marginal.
In the Khanag subdialect, the "horizontal" system can be more complicated, being extended with the third member tu. Uslar describes the Khanag "horizontal" system as ternary: mu 'this (near the speaker)' / du-mˈu 'this (near the addressee)' / tu-mˈu 'that (far from the speaker and addressee)'. The analysis is actually not quite correct, as follows from Uslar's own remark: "This explanation of mu / du-mˈu / tu-mˈu is undoubtedly right, because it has been adopted from various sources. But Tabasaran speakers themselves pay little attention to such semantic nuances in the natural speech" [Uslar 1979: 136]. Apparently Uslar implies that the opposition between du-mˈu and tu-mˈu is in fact desemanticized.
According to [Dirr 1905: 37 f.], however, the Khanag "horizontal" system is binary: mu (~ mu-mu) 'this' / du-mu (~ tu-mu ~ du ~ tu) 'that'. See Common Tabasaran notes on this controversy.
As described in [Kibrik et al. 1977a 2: 124] and [Kibrik 1994: 319], the system of Archi demonstrative attributive pronouns on the horizontal axis (i.e., excluding the vertical oriented forms) is ternary: ya-class 'this (near the speaker)' / yˈa-mu-class 'this (near the addressee)' / to-class 'that (far from the speaker and addressee)'. According to [Dirr 1908: 29], ya-class / to-class are statistically the most frequently used demonstrative pronouns. We prefer to assume that the basic Archi system is binary: ya-class 'this' (near the speaker)' / to-class 'that (far from the speaker and addressee)'; in turn, the complex demonstrative yˈa-mu-class 'this (near the addressee)' looks like a recent formation, created for some specific semantic or pragmatic nuances.
Two additional "vertical" members are gu-dˈu 'that (below)', ʁu-dˈu 'that (above)' [Kibrik et al. 1977a 2: 124; Kibrik 1994: 319].
According to Magometov's data, the Khiv system of attributive deictic pronouns on the horizontal axis is binary: mu 'this' / du-mu (~ ha-t-mu) 'that'. On the variant with -t- (ha-t-mu) see Common Tabasaran notes. Two Khiv "vertical" pronouns are: kːu-mu 'that (below)' / ʁu-mu 'that (above)' [Magometov 1965: 178].
The Literary Tabasaran "horizontal" system is also binary: mu 'this' / du-mu 'that', as described in [Zhirkov 1948: 98 ff.]. According to Zhirkov, "vertical" pronouns 'that (below)' / 'that (above)' were lost. On the contrary, in [Khanmagomedov & Shalbuzov 2001: 151, 282, 433] the literary "horizontal" system is evasively described as mu 'this' / t-mu 'that' with the third member du-mu 'that' (difference between t-mu and du-mu is not explicated).
It is possible that the Proto-Tabasaran system of attributive deictic pronouns on the horizontal axis was indeed ternary, as it is described by Uslar for the Khanag subdialect: *mu 'this (near the speaker)' / *du 'that (near the addressee)' / *tu 'that (far from the speaker and addressee)'. In all the dialects, the absolutive forms of the two latter members are secondarily modified with the desemanticized morpheme -mu, i.e. du-mu and tu-mu.
In fact, however, the third member tu (or tu-mu) seems very rare and marginal in the described Tabasaran dialects. In [Dirr 1905: 37 ff.], the forms tu and tu-mu 'that' is noted only in parentheses as a variant of du(-mu); in [Magometov 1965: 176 ff.], only the emphatic form ha-t-mu 'exactly that' is observed; no forms with the morpheme t(u) are quoted in [Zhirkov 1948: 98 ff.].
Thus, according to the dialectal overview in [Magometov 1965: 176 ff.], the common Tabasaran basic opposition on the horizontal axis is currently binary: mu 'this' / du-mu 'that'.
The Literary Lezgi system has been semantically transformed: i 'this' / a 'that' / a-tʼˈa 'yonder' / a-ʁˈa 'that (below)' / wi-nˈi 'that (above)' [Alekseev & Sheykhov 1997: 46; Haspelmath 1993: 190] (somewhat differently in [Gaydarov et al. 2009: 156]). As noted in [Haspelmath 1993: 190], however, only i 'this' and a 'that' are frequent in the modern literary language, other pronouns are marginal. Thus the Literary Lezgi system is, in fact, binary.
NCED: 993. Distribution: The Proto-Lezgian system of the demonstrative pronouns can hardly be reconstructed in all details. It is particularly unclear whether the system was binary - 'this' / 'that' - or ternary: 'this' / 'that (near)' / that (far)'. Nevertheless, some general considerations can be proposed. The main data on the pronouns on the horizontal axis are summarized as follows (synchronically basic forms are bold-faced as opposed to statistically marginal ones):
| 'THIS/THAT' | Udi | Archi | Kryts | Budukh | Tsakhur | Rutul | Aghul | Tabasaran | Lezgi |
| *mV [NCED: 842] | me this | ya-mu that (near) | a-m that (near) | ma this | mi this | me this | mu this | ||
| *tʼV | tːe that (far) | tʼa that (far) | |||||||
| *tV [NCED: 993] | to that (far) | ti that | ti that | tu that | |||||
| *dV [NCED: 404] | du that | ||||||||
| *kV | ka that (near) | ||||||||
| *ʔi [NCED: 214] | y-a this | i this | i this | ||||||
| *lV [NCED: 775] | li this lä that | u-lu this a-la that (far) | |||||||
| *šV- [LEDb: #187] | še that | ||||||||
| *ha- [NCED: 486] | ha that | ||||||||
| *ʔa [NCED: 218] | a-la that (far) | a that (near) | |||||||
| *ʔu [NCED: 222] | u-lu this |
The demonstrative pronoun of proximal deixis 'this' can be assuredly reconstructed as *mV [NCED: 842]: this function was retained in Udi and many Nuclear Lezgian languages (Tsakhur, Rutul, Aghul, Tabasaran). In Archi, this root forms the secondary pronoun of medial deixis ya-mu 'near the addressee'.
Reconstruction of the distal deixis pronoun 'that' is more complicated. As one can see, forms that originate from virtual *tʼV and virtual *tV are in complementary distribution among languages. Thus, it is reasonable to suppose that these go back to a single protoform with occasional sound irregularities (this is not rare in grammatical morphemes). External comparison suggests that the original shape should be *tV [NCED: 993]; if so, we deal with secondary tenseness in Udi and secondary ejectivization in Lezgi (both cases are probably iconically conditioned). The pronoun *tV (~ *tʼV) is the obvious candidate for the status of the Proto-Lezgian pronoun 'that' (this function was retained in Udi, Archi and the most of Nuclear Lezgian lects).
It should be noted that in [NCED: 404], the ejective forms in Udi (tːe) and in Lezgi (tʼa) are included into the entry *dV. First, this is not likely due to the aforementioned considerations. Second, evidence for the Proto-Lezgian demonstrative morpheme *dV is relatively weak: this is the basic Tabasaran pronoun du 'that' and the second element of the Archi vertical pronoun gu-dˈu 'that (below)', ʁu-dˈu 'that (above)'. Other descendants of *dV, listed in [NCED: 404], should be excluded: Kryts ä-d, Budukh a-d 'that' (used independently), where -d is in fact a synchronic class exponent; apparently in Rutul ha ~ ha-d '(s)he, it, they', the final -d represents the same fossilized grammatical morpheme.
Note also that it is theoretically possible to reconstruct *tV (~*tʼV) as the specific pronoun of distal deixis 'that (far)'. In such a case, *ʔi [NCED: 214] is the main candidate for the Proto-Lezgian pronoun of medial deixis 'that (near)'.
To sum up, the Proto-Lezgian system of the demonstrative pronouns in the horizontal row can be reconstructed as binary *mV 'this' / *tV 'that' or (with less certainty) ternary *mV 'this' / *ʔi 'that (near)' / *tV that (far)'.
The totally restructured South Lezgian (Kryts, Budukh) system probably originates from the old "vertical" demonstrative pronouns, thus in [NCED: 775].
The morpheme *ha- [NCED: 486] is the Common Lezgian pronominal emphatic "augment" (thus Rutul ha 'that' may in fact originate from *ha-ʔa). The original Proto-Lezgian status and exact function of other demostrative morphemes listed above are not clear.
Reconstruction shape: Correspondences are regular except for (iconic?) ejectivization/tenseness in Udi and Lezgi and vowel fluctuation among various lects (perhaps of contracted origin).
Semantics and structure: Primary pronominal root.
Caucasian Albanian: see notes on 'that'.
Tsakhur-Kum Tsakhur: see notes on 'that'.
Reconstruction shape: Correspondences are regular, except for vowel fluctuation among various lects.
Semantics and structure: Primary pronominal root.
Caucasian Albanian: vun [abs., erg.] / veː [gen.] / va- [obl.] [Gippert et al. 2008: II-37, IV-39].
Tsakhur-Kum Tsakhur: ʁu [Kibrik & Kodzasov 1990: 222; Schulze 1997: 37]. Paradigm: ʁu [abs., erg.] / yiʁ-ɨn [gen.].
Muxrek dialect: wɨ [abs.] / wɨ-dɨ [gen.] / wa-d [erg.] / wa-s [dat.] [Ibragimov 1978: 177].
Borch-Khnov dialect: ʁu [abs.] / ʁu-du [gen.] / ʁu-ǯ [erg.] / ʁʷa-s [dat.] [Ibragimov 1978: 262].
One of the most economic scenarios presupposes reconstructing for Proto-Rutul a suppletive paradigm of 'thou' with four stems: *wɨ [abs.] / *ʁu [erg.] / *yuʁ-dɨ [gen.] / wa- [dat.]. In Mukhad, Muxrek & Luchek the paradigm was completely restructured after the absolutive and oblique stems wɨ / wa-: gen. wɨ-dɨ is a regular synchronic formation on the basis of abs. wɨ, whereas the new ergative form is based on oblique wa-. In Borch-Khnov the paradigm, on the contrary, was analogically levelled after the ergative ʁu; oblique ʁʷa- represents a hybrid of old *wa- and the new stem ʁu-. Similarly, the Ixrek paradigm was levelled up after ergative ʁu, but the old genitive yuʁ-dɨ is still retained. The situation in Shinaz is more complicated: apparently the old genitive form was superseded with the new formation based on abs. wɨ (an innovation shared with the Mukhad, Muxrek & Luchek dialects, which surround the Shinaz area, see the map in [Ibragimov 1978: 14]) and subsequently Shinaz abs. wɨ was superseded with the ergative morpheme ʁu.
The same in the Khudig subdialect: wu-n [abs., erg.] / wa-s [dat.] / yew [gen.] 'thou' [Magometov 1970: 101].
The same in the Khanag subdialect: iwˈu [abs., erg., obl.] / yaw [gen.] 'thou' [Uslar 1979: 126; Dirr 1905: 33; Magometov 1965: 170].
The same in the Khiv subdialect: uwˈu [abs., erg., obl.] / yaw [gen.] 'thou' [Magometov 1965: 169]. The same in Literary Tabasaran: uwˈu [abs., erg., obl.] / yaw [gen.] {уву} 'thou' [Khanmagomedov & Shalbuzov 2001: 432; Zhirkov 1948: 107; Alekseev & Shikhalieva 2003: 56].
The same in Literary Lezgi: wu-n [abs.] / wu-n-ˈa ~ na [erg.] / wi [gen.] / wa- [obl.] {вун} 'thou' [Gaydarov et al. 2009: 150; Haspelmath 1993: 184].
The same in the Akhty dialect: Khlyut wɨ-n [abs.] / wɨ-n-ˈa [erg.] / wi [gen.] / wa- [obl.] 'thou' [Kibrik & Kodzasov 1990: 222].
In almost all the lects (except for Tsakhur, Rutul), ʁ-forms were completely eliminated. The new ergative forms coincide with abs. *u̯o(-n) or, occasionally, are based on obl. *u̯a-. The new genitive forms normally represent something like *u̯i- or *u̯e-, although in Kryts, the genitive is based on obl. *u̯a-, whereas Koshan Aghul and Tabasaran yaw ~ yew 'of thee' is a hybrid of old *class=oʁ(ʷ) and the u̯-forms. It should be noted that the origin of the common genitive *u̯i- ~ *u̯e- is unclear.
The only group in which ʁ-forms have survived is West Lezgian (Tsakhur, Rutul). In Tsakhur, the absolutive coincides with the ergative *ʁu. Many Tsakhur dialects, however, demonstrate the subsequent secondary derivation of the ergative and genitive forms from obl. *u̯a-. The Rutul situation is more complicated, although it is possible to derive all the variety of attested Rutul systems from the Proto-Rutul paradigm *wɨ [abs.] / *ʁu [erg.] / *yuʁ-dɨ [gen.] / wa- [dat.], which directly continues the assumed Proto-Lezgian paradigm.
Note that pace [NCED: 483] Nidzh Udi hun has nothing to do with *ʁu-n, but originates from Proto-Caucasian Albanian-Udi *wu-n with the (not infrequent) Nidzh prothesis h-.
As in the case of the 1st sg. pronoun 'I' q.v., the Tabasaran absolutive-ergative-oblique stem iwu- represents a hybrid of the old absolutive and genitive forms.
The absolutive form can be safely reconstructed with the nasal suffix: *u̯o-n. The only lects that lack the nasal element are some Rutul dialects (wɨ for expected **wɨ-n), where we probably deal with influence on the part of erg. ʁu 'thou' and abs. zɨ 'I'.
Reconstruction shape: Basic correspondences seem regular, although *u̯ is a very rare Proto-Lezgian phoneme.
Semantics and structure: Primary pronominal root. The suppletive paradigm: *u̯o-n [abs.] / *ʁu [erg.] / *class=oʁ(ʷ) [gen.] / *u̯a- [obl.].
Caucasian Albanian: muz 'tongue; language' [Gippert et al. 2008: IV-30].
Tsakhur-Kum Tsakhur: miz [Kibrik & Kodzasov 1990: 15].
The same in the Khanag subdialect: milʒ 'tongue' [Uslar 1979: 856, 1010; Dirr 1905: 195, 247]. The same in other subdialects: Khyuryuk, Kumi milʒ {милзз} 'tongue' [Genko 2005: 120].
The same in the Khiv subdialect: melz {мелз} 'tongue' [Genko 2005: 119]. The same in Literary Tabasaran: melz {мелз} 'tongue' [Khanmagomedov & Shalbuzov 2001: 227].
The same in Literary Lezgi: mez [abs.] / mecː-ˈi- [obl.] {мез} 'tongue' [Talibov & Gadzhiev 1966: 233; Gadzhiev 1950: 962; Haspelmath 1993: 498, 528].
The same in the Akhty dialect: Khlyut mez [abs.] / mecː-ˈa- [obl.] 'tongue' [Kibrik & Kodzasov 1990: 15].
Replacements: {'tongue' > 'language'} (Caucasian Albanian, Udi, Alyk Kryts, Tsakhur).
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Primary substantive root. The oblique stem is *melcːe-.
Distinct from Nidzh meran ‘fang’ [Kibrik & Kodzasov 1990: 17]. In [Gukasyan 1974: 173] quoted as merun ‘fang of wild boar’ (according to Gukasyan, attested in both Nidzh and Vartashen). A term of unclear origin.
Caucasian Albanian: Unattested.
Distinct from the specific term gˈaˤnži 'molar' [Kibrik & Kodzasov 1990: 17; Kibrik et al. 1977b: 230; Chumakina et al. 2007; Dirr 1908: 137].
Distinct from the specific term saχ 'molar' [Kibrik & Kodzasov 1990: 17].
Distinct from the specific term azu ~ azu sil 'molar' [Kibrik & Kodzasov 1990: 17; Meylanova 1984: 18], borrowed from Azerbaijani azɨ diši 'molar' (Budukh phonetics suggests the dialectal variant *azu).
Distinct from azɨ 'molar' [Ibragimov 1990: 149; Ibragimov & Nurmamedov 2010: 29], borrowed from Azerbaijani azɨ diši 'molar'.
Tsakhur-Kum Tsakhur: sʸilʸi [Kibrik & Kodzasov 1990: 16]. Distinct from the specific term azɨ 'molar' [Kibrik & Kodzasov 1990: 17], borrowed from Azerbaijani azɨ diši 'molar'.
A second candidate is sɨs {сыс} 'tooth' [Dzhamalov & Semedov 2006: 236, 343]; no examples have been found, but sɨs is used in the expression for 'to bite' q.v.: sɨs haʔ-, literally 'tooth + to do'.
Distinct from the more specific term wɨgːɨn {выггын} with polysemy: 'fang / wedge' [Dzhamalov & Semedov 2006: 49].
Distinct from the more specific terms: ʁʷab-ad sɨs 'molar', literally 'tooth of root' [Kibrik & Kodzasov 1990: 17] and wɨgɨn 'fang' [Kibrik & Kodzasov 1990: 17].
It seems that the Proto-Rutul term for 'tooth (in general)' should be reconstructed as sɨl; early on, its plural form sɨl-ab spread onto the singular number in all dialects, but currently the new singular form sɨl-ab is being superseded by the unrelated form sɨs. The underlying meaning of sɨs can be 'incisor' or 'canine tooth', because in many Rutul dialects the expression for 'to bite' q.v. is based on this substantive: sɨs + 'to do'.
Distinct from seχʷ 'molar' [Kibrik & Kodzasov 1990: 17] and kːančː 'fang' [Kibrik & Kodzasov 1990: 17].
The same in the Khanag subdialect: silˈib [abs.] / silb-ˈ [obl.] 'tooth' [Uslar 1979: 908, 994]; in [Dirr 1905: 204, 230], quoted as selew [abs.] / selb- [obl.] 'tooth' with additional phonetic development b# > w#. Distinct from Khanag šarʁ or šarʁˤ 'molar' [Dirr 1905: 220].
The same in the Khyuryuk subdialect: silib {силиб} 'tooth' [Genko 2005: 139]; distinct from Khyuryuk šarʁ {шаргъ} 'molar' [Genko 2005: 189] (apparently Genko's error for expected **šarʁˤ {шяргъ}).
The same in the Khoredzh and Chara subdialects: slib [abs.] / spː- [obl.] 'tooth' [Magometov 1965: 58]; distinct from Chara šarʁ {шаргъ} 'molar' [Genko 2005: 189] (apparently Genko's error for expected **šarʁˤ {шяргъ}).
The same in Literary Tabasaran: silˈib [abs.] / silb-ˈ [obl.] {силиб} 'tooth' [Khanmagomedov & Shalbuzov 2001: 271]. Distinct from two literary terms for 'molar': saχʷ {сахв} 'molar' [Khanmagomedov & Shalbuzov 2001: 270] and šarʁˤ {шяргъ} 'molar' [Khanmagomedov & Shalbuzov 2001: 353]. Distinct from two literary terms for 'fang': kːanč {кканч} with polysemy: 'fang (of wild boar) / high temple (hair)' [Khanmagomedov & Shalbuzov 2001: 196] (the primary meaning is 'fang') and kʼir {кIир} with polysemy: 'hook / fang' [Khanmagomedov & Shalbuzov 2001: 214] (the primary meaning is 'hook').
Differently in the Khiv subdialect: sars {сарс} 'tooth' [Genko 2005: 137]. Distinct from Khiv saχʷ {сахв} 'molar' [Genko 2005: 138] and kːanč {кканч} 'fang (of wild boar)' [Genko 2005: 92].
Distinct from Gyune sʷaχ [abs.] / suχʷ-ˈa- [obl.] 'molar' [Uslar 1896: 546, 613]. Cf. also Gyune kʼir, which is glossed only as 'hook' in [Uslar 1896: 477].
The same generic term in Literary Lezgi: sas [abs.] / sar-ˈa- [obl.] / sar-ˈar [pl.] {сас} 'tooth' [Talibov & Gadzhiev 1966: 287; Gadzhiev 1950: 250; Haspelmath 1993: 505, 528] (incorrectly glossed as 'incisor' by Haspelmath). Distinct from literary sʷaχ [abs.] / suχʷ-ˈa- [obl.] {свах} 'molar' [Talibov & Gadzhiev 1966: 288; Haspelmath 1993: 506], kʼir {кIир} 'hook / fang' [Talibov & Gadzhiev 1966: 217] and gung {гунг} 'fang (of wild boar)' [Talibov & Gadzhiev 1966: 89].
The same in the Akhty dialect: Khlyut sas [abs.] / ss-a- [obl.] / ss-ar [pl.] 'tooth' [Kibrik & Kodzasov 1990: 16]. Distinct from Khlyut kːʷalˈaχ 'molar' and kʼir 'fang' [Kibrik & Kodzasov 1990: 17].
In some Rutul dialects (Luchek, Khnyukh, Shinaz), Southern Tabasaran (Khiv) and probably in all Lezgi dialects, this stem was superseded with *sars (~ sː) [LEDb: #110] - an areal Lezgi-induced isogloss. The original meaning of *sars is unclear, because this root is not attested outside Rutul, Tabasaran and Lezgi and lacks external North Caucasian etymology (some Rutul data may point to the local meaning 'incisor' or 'canine tooth', although it is not certain).
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Primary substantive root. In Udi and many Nuclear Lezgian lects the old plural forms, modified with fossilized suffixes, are currently used in the singular meaning. Due to this reason, the oblique stem is not reconstructible.
Caucasian Albanian: Unattested. The restoration of the lexeme χod 'tree' is uncertain, see [Gippert et al. 2008: IV-22, VII-21, 90 fn. 52]. Cf. the loanword durud 'piece of wood' [Gippert et al. 2008: IV-13], for which see above.
Distinct from cʼaħˈanː- 'wood, timber' [Kibrik et al. 1977b: 208; Chumakina et al. 2007; Comrie & Khalilov 2010: 49] (in the light of ħ should be a loanword).
Distinct from kʼɨnä 'log, firewood' [Kibrik & Kodzasov 1990: 162].
Distinct from kʼɨnǝ {кIына} 'log, (piece of) wood, firewood' [Meylanova 1984: 99; Kibrik & Kodzasov 1990: 162].
Distinct from os 'log, firewood' [Kibrik et al. 1999: 883].
Tsakhur-Kum Tsakhur: yiw 'tree', os 'log, firewood' [Kibrik & Kodzasov 1990: 95, 162].
Distinct from os 'log, firewood' [Kibrik & Kodzasov 1990: 162].
Distinct from os 'log, firewood' [Kibrik & Kodzasov 1990: 162].
Distinct from us '(fire)wood' [Dirr 1912: 175].
Distinct from us 'log, (piece of) wood, firewood' [Dzhamalov & Semedov 2006: 256].
Distinct from us 'log, firewood' [Kibrik & Kodzasov 1990: 162].
Distinct from Burshag aχʷ 'forest' [Kibrik & Kodzasov 1990: 95; Shaumyan 1941: 168].
Distinct from Richa dar 'forest' [Kibrik & Kodzasov 1990: 95].
The same opposition in the Usug subdialect: kʼur 'tree', dar 'forest' [Shaumyan 1941: 168, 184].
Distinct from dar with polysemy: '(piece of) wood / forest' [Dirr 1907: 112; Kibrik & Kodzasov 1990: 95; Shaumyan 1941: 168].
Distinct from Tpig dar 'forest; cudgel' [Suleymanov 2003: 74; Shaumyan 1941: 168].
The same in the Khanag subdialect: har with polysemy: 'tree / forest' [Uslar 1979: 708, 992; Dirr 1905: 172, 227]. Distinct from Khanag čʼerˈeʔ 'small tree' [Uslar 1979: 973; Dirr 1905: 219] and kʼakʼˈul 'wood, cut down tree, log, firewood' [Uslar 1979: 794; Dirr 1905: 189].
The same in the Khyuryuk subdialect: har {гьар} with polysemy: 'tree / forest' [Genko 2005: 50]. Distinct from Khyuryuk čʼerˈeʔ {чIереъ} 'small tree' [Genko 2005: 185] and kʼakʼˈul {кIакIул} 'log, cut down tree' [Genko 2005: 110].
The same in the Khiv subdialect: har {гьар} with polysemy: 'tree / forest' [Genko 2005: 50]. Distinct from Khiv wergˈil {вергил} 'small tree' [Genko 2005: 32] and gakʼˈul {гакIул} 'wood, log, firewood' [Genko 2005: 35]. For the dissimilative phonetics of the latter term cf. kʼakʼˈul {кIакIул} 'log, cut down tree' in the Tinit subdialect [Genko 2005: 110], as well as the Northern Tabasaran forms above.
Similarly in Literary Tabasaran: har {гьар} with polysemy: 'tree / pole' [Khanmagomedov & Shalbuzov 2001: 125]. Distinct from literary gakʼˈul {гакIул} 'wood, firewood' [Khanmagomedov & Shalbuzov 2001: 95] and yarkˈur {яркур} 'forest, grove' [Khanmagomedov & Shalbuzov 2001: 361].
The same in Literary Lezgi: tːar {тар, ттар} with polysemy: 'tree / wooden bar (e.g., of door)' [Talibov & Gadzhiev 1966: 306; Gadzhiev 1950: 167; Haspelmath 1993: 507, 528]. Distinct from literary kʼarˈas {кIарас} 'wood, firewood' [Talibov & Gadzhiev 1966: 212; Gadzhiev 1950: 167; Haspelmath 1993: 496] and čχa {чха} 'split billet' [Talibov & Gadzhiev 1966: 376].
The same in the Akhty dialect: Khlyut tːar 'tree' [Kibrik & Kodzasov 1990: 95]. Distinct from Khlyut kʼarˈas 'wood, firewood', čχay 'log, billet' [Kibrik & Kodzasov 1990: 162].
The second candidate is *χːʷar ~ *χːʷar-tː [NCED: 1079] (as proposed in [NCED: 888], Lezgian *-tː < North Caucasian *-di is a suffix, which sometimes modifies plant names). The prefixed stem *χːʷar-tː means 'tree' in Udi, whereas in other Lezgian languages, *χːʷar and *χːʷar-tː denote 'log, pole', 'ceiling', 'beam'. External North Caucasian comparanda (if correct) suggest that the original meaning of Lezgian *χːʷar ~ *χːʷar-tː could be 'a k. of foliage tree' (probably an endemic species, which was basically used as building wood). It must be noted that pace [NCED: 1079], Luchek Rutul χɨd 'lime tree' is not related here, but regularly originates from Proto-Lezgian *χer / *χer-tː 'lime tree' [NCED: 888].
The third candidate is *tːar [NCED: 399]. This root is attested with the meaning 'tree' in South Lezgian (Kryts Proper, Budukh) and Lezgi, having been lost in the rest of languages (except for Aghul, see below). Formal distribution suggests that this item should be reconstructed as the Proto-Nuclear Lezgian term for 'tree'. In non-Koshan Aghul, *tːar means 'forest', except for Fite Aghul, where dar has secondarily acquired the polysemy 'tree / forest'. It is theoretically possible to suppose that dar actually represents the Proto-Aghul term for 'tree' and the shift 'tree' > 'forest' for dar is a late areal isogloss, which has not fully affected the Fite border territory. Such a solution does not seem apt, however, because the Common Aghul root for 'tree' is *kʼʷir, attested with this meaning in all Aghul dialects including Fite (there are two synchronic synonyms for 'tree' in Fite).
Several replacements of *tːar occurred in individual Nuclear Lezgian lects. The most interesting substitution took place in Alyk Kryts, where 'tree / (piece of) wood' is expressed with the root *kʼon-(ay) [NCED: 727]. The Proto-South Lezgian meaning of *kʼon-(ay) was '(piece of) wood, firewood' (as follows from the Kryts Proper and Budukh cognates), whereas the Archi and external North Caucasian comparanda suggest that the original Proto-Lezgian meaning of this root should be 'handle, grip'.
In Rutul, *tːar was superseded with *χːʷɨ(r)k [NCED: 1082]. The original meaning of the latter root should be 'forest' as follows from its Archi cognate χːʷak 'forest' and the polysemy 'tree / forest' in Muxrek Rutul.
In Aghul, *tːar was superseded with *kʼʷir (~ -ɨ-) [LEDb: #139]. This is a Common Lezgian root, attested in both Udi and some Nuclear Lezgian lects, where it denotes 'wooden pole', 'ladder-step', 'roofing wood'. The exact original meaning of *kʼʷir (~ -ɨ-) is not reconstructible.
Unetymologizable equivalents for 'tree' are attested in Tsakhur (yiw), Tabasaran (har).
Replacements: {'(piece of) wood, firewood' > 'tree'} (Alyk Kryts), {'forest' > 'tree'} (Rutul, Fite Aghul), {'a k. of foliage tree' > 'tree'} (Udi), {'tree' > 'tree stump, stub'} (Luchek Rutul), {'tree' > 'forest'} (non-Koshan Aghul).
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Primary substantive root. The Archi form is modified with the common desemanticized nominal suffix -*Vy.
Caucasian Albanian: pʼʕa [Gippert et al. 2008: II-40, IV-36].
Tsakhur-Kum Tsakhur: qʼˤo-class-lʸe [Kibrik & Kodzasov 1990: 247].
The same in the other subdialects: Tsirkhe qʼˤu-d, Khpyuk ʡu-d 'two' [Suleymanov 1993: 119].
The same in the Khanag subdialect: qʼˤu-class 'two' [Uslar 1979: 151; Dirr 1905: 42]. The same in the Khyuryuk subdialect: qʼˤu-class {кьюв} 'two' [Genko 2005: 108].
The same in the Khiv subdialect: qʼˤu-class {кьюб} 'two' [Genko 2005: 107]. The same in Literary Tabasaran: qʼˤu-class {кьюб} 'two' [Khanmagomedov & Shalbuzov 2001: 210; Zhirkov 1948: 91; Alekseev & Shikhalieva 2003: 52 f.].
The same in Literary Lezgi: qʼʷe {кьве} 'two' [Talibov & Gadzhiev 1966: 566; Haspelmath 1993: 230] (in the non-attributive function, the variant qʼʷe-d {кьве} is used).
The same in the Akhty dialect: Khlyut qʼʷä-d 'two' [Kibrik & Kodzasov 1990: 247].
Final -d is the old class exponent.
Reconstruction shape: Correspondences seem regular.
Semantics and structure: Primary numeral root.
Udi verbs of motion (particularly basic 'to come' q.v. and 'to go') represent a certain morphological riddle. As discussed in [Harris 2002: 68 ff., 223 ff.; Maisak 2008a: 107 ff., 154 ff.] and some other publications and now strongly suggested by data from Caucasian Albanian, the attested Udi paradigm originates from the preverb ta- 'thither, away' plus the verb 'to go', which is reflected as Caucasian Albanian iʁ- (present-infinitive) / ac- (past) / ukal- (imperative); that is, Proto-Udi *ta=iʁ- (present-infinitive) / *ta=ac- (past) / *ta=iʁ- (future) / *ta=k- (imperative).
The attested Udi present stem ta(y)- goes back to *ta=iʁ- with ʁ > 0/y in the intervocalic position before front vowels or before consonants, cf. [Schulze 2005: 547 (3.4.2.1 #34)]. Accordingly, as it was correctly suspected in [Schulze-Fürhoff 1994: 474] and [Harris 2002: 223 ff.], the Udi present ta=y- contains the same etymological morpheme as the future ta=ʁ-. It must be noted that in ta(y)- 'to go' and in another prefixed verb from the same root, e=y- 'to come' q.v., this -ʁ- is still sporadically retained in present and masdar forms in both Udi dialects (taʁ-esun etc.) [Schulze 2005: 547 (3.4.2.1 #34); Fähnrich 1999: 15, 30]. Alternatively it is supposed in [Maisak 2008a: 107, 145 f.] that the present stems ta=y- 'to go' and e=y- 'to come' originate from the imperative ta=k-, e=k- with k > y, although such a phonetic development is not regular in the forms discussed (the only known instance of such a shift is the verb bak-es 'to be' [Maisak 2008a: 144]).
The preverb ta- 'thither, away' is attested in few verbs in modern Udi (ta(y)- 'to go', taš- 'to carry away' and tad- 'to give' q.v., see [Maisak 2008a: 158]), but possesses a slightly broader distribution in Caucasian Albanian [Gippert et al 2008: II-45].
The Udi and Caucasian Albanian suppletive verbs 'to come' q.v., Udi e(y)- (present-infinitive) / ar- (past) / eʁ- (future) / ek- (imperative), Caucasian Albanian heʁ- (present-infinitive) / ar- (past) / hekal- (imperative), represent an almost exact morphological parallel to 'to go'. As correctly analyzed in [Schulze 2005: 547 (3.4.2.1 #34); Gippert et al. 2008: II-45, 51], verbs for 'to come' contain the preverb (h)e- 'hither' plus the plain verb 'to go'. Thus, Udi *e=iʁ- > *e=y- (present-infinitive) / *e=iʁ- (future) / *e=k- (imperative) 'to come' and Caucasian Albanian *he=iʁ- (present-infinitive) / he=kal- (imperative) 'to come'.
The verbal prefix (h)e- 'hither' is very scantily preserved in modern Udi [Maisak 2008a: 158], but is more widely attested in Caucasian Albanian [Gippert et al 2008: II-45].
Some authors (see discussions in [Harris 2002: 68 ff., 223 ff.] and [Maisak 2008a: 107 ff., 154 ff.]) suppose that the original root for 'to come' was *eʁ- / *ek-; afterwards, the initial e- was morphologically reanalyzed in Pre-Udi as a verbal prefix e- with the new meaning 'hither'. Data from Caucasian Albanian, where the preverb (h)e- is more productive than in the modern language, contradicts such an analysis.
Thus, we believe that the following present-infinitive-future (scil. imperfective) stems can assuredly be reconstructed for Proto-Caucasian Albanian-Udi:
1) *iʁ- 'to go' (retained in CA, lost in Udi);
2) *ta=iʁ- 'to go away (to go thither)' (retained in CA, broadened into 'to go / to go away' in Udi);
3) *(h)e=iʁ- 'to come (to go hither)' (retained both in CA and Udi).
Modern Udi demonstrates the typologically normal situation, in which the meanings 'to go (in general)' and 'to go away' merge into one verbal root, which is lexically opposed to an expression for 'to come'.
As for the past (scil. perfective) stem, here the opposition between two main concepts 'to go, go away' and 'to come' is expressed by different roots:
1) *ac- 'to go' (retained in CA, lost in Udi);
2) *ta=ac- 'to go away' (retained in CA, broadened into 'to go / to go away' in Udi);
3) *ar- 'to come' (retained both in CA and Udi).
Note that in fact the plain *ac- survived in Modern Udi as the past stem -(e)c- of the light verb -e-, which functions as an intransitivizer or decausative (for which see [Maisak 2008a: 139 ff.; Schulze 2005: 562 ff. (3.4.2.2 #11 ff.)]).
The situation with the imperative stems is less clear etymologically. Cf. Caucasian Albanian u(=)kal- ~ e=k-e- 'to go (imv.)', he=kal- 'to come (imv.)' [Gippert et al 2008: II-51] vs. Udi ta=k-e 'to go (imv.)', e=k-e 'to come (imv.)' [Schulze 2005: 552 (3.4.2.1 #44)]. Schulze & Gippert might theoretically be right that Udi =k- is an allegro variant of *=kal- [Gippert et al. 2008: II-51]. At first sight, the CA form ukal- looks like u=kal- with a unique prefix u- [Gippert et al. 2008: II-51], but the initial u can also be a normal reflex of the first vowel of the proto-root, see notes on up- 'to say'.
Quite a different analysis has been proposed in [NCED: 423] (whose authors were naturally not aware of Causasian Albanian data): Udi ta(y)- 'to go, to go away' is explained as a secondary formation from the unattested Udi gerund *atay and further compared to Lezgi at- 'to come, arrive', which does not currently seem apt.
Caucasian Albanian: A suppletive verb iʁ- (present-infinitive) / ac-e- (past) / ukal- (imperative) [Gippert et al 2008: II-44, 45, 51, IV-21]. Also attested with the preverb ta- 'thither': ta-iʁ-esun 'to go away, thither' [Gippert et al 2008: II-45].
Distinct from a more specific verb =ˈerqˤa- 'to walk, go' [Chumakina et al. 2007; Kibrik et al. 1977b: 239; Kibrik et al. 1977a 3: 235; Kibrik & Kodzasov 1988: 73; Comrie & Khalilov 2010: 491; Mikailov 1967: 178]. Both verbs are probably etymologically related, representing various grades of Proto-Lezgian Ablaut [NCED: 572 f.]; the medial -r- in =ˈerqˤa- 'to walk' is a fossilized durative infix; initial h-, which appears in the IV class forms of =ˈerqˤa- (hˈerqˤa-), is not entirely clear.
The same Proto-Lezgian root is retained as the perfective stem =qˤˈa of ‘to come’ q.v.
From the etymological point of view, it seems probable that imperfective čä-r- originates from *č=äʁ-r- (cf. its Budukh counterpart), although the loss of ʁ seems irregular; in this case č= is the prefix 'next to' [Saadiev 1994: 424]. Perfective =ix- is the same root as 'to come' q.v.
According to [Kibrik et al. 1999], there are two suppletive verbs, both of them with polysemy: 'to go / to go away':
1) aˤlʸ=haː- < *aˤlʸ=ha-a- [imperf., fut.] / akʼɨn [perf.] / h=oːr-a [imv.] 'to go; to go away; to depart' [Kibrik et al. 1999: 69, 868; Ibragimov & Nurmamedov 2010: 35]. In addition, the ablaut variant akʼan- is used in imperf. & fut. with the iterative meaning 'to go away constantly'.
2) ɨqːˤ-a- ~ uqːˤ-a- [imperf., fut.] / h=akʼɨn [perf.] / h=akʼnʸ-e [imv.] 'to go; to go away' [Kibrik et al. 1999: 879; Ibragimov & Nurmamedov 2010: 135, 355]. In addition, the ablaut variant h=akʼan- is used in imperf. with the iterative meaning 'to go constantly; to go away constantly'.
Browsing through texts in [Kibrik et al. 1999] does not permit to make a definite choice between these two verbs, so we are forced to treat them as synonyms, including both imperfective and perfective stems (that is, three synonymous items are proposed for this slot: aˤlʸ=ha-, akʼɨn ~ h=akʼɨn, ɨqːˤ- ~ uqːˤ-).
Distinct from the ablaut verb h=ohar- [imperf., fut.] / h=ohur- [perf.] 'to walk' [Kibrik et al. 1999: 878].
According to [Kibrik & Kodzasov 1988], the situation is similar to Mishlesh: there are two synonymous verbs for 'to go', one of them with polysemy: 'to go / to go away':
1) aˤl=h-a- [imperf., fut.] 'to go; to go away' / akʼɨn 'to go away' [perf.] / h=oːr-a [imv.] 'to go' / akʼnʸ-e [imv.] 'to go away' [Kibrik & Kodzasov 1988: 73, 74; Dirr 1913: 136, 164, 176].
2) ɨqːˤ-a- [imperf., fut.] / h=akʼɨn [perf.] / h=akʼnʸ-e [imv.] 'to go' [Kibrik & Kodzasov 1988: 73; Dirr 1913: 163].
We are forced to treat aˤl=h-, h=akʼɨn and ɨqːˤ- as synonyms.
According to the data in [Makhmudova 2001] and [Ibragimov 1978], the suppletive paradigm is as follows: class=iʔi ~ class=r=uʔ-u-r- [imperf.] / class=ɨx-ɨ-r [perf.] / class=rɨχ [imv.] / ma=class=ʁʷˤ [prohib.] / y=ix-a-s [fut.]. In imperf., with polysemy: 'to go / to come'.
According to the data in [Dzhamalov & Semedov 2006], the suppletive paradigm is as follows: class=iʔi ~ class=r=uqʼˤ-u-r- [imperf.] / class=ɨx-ɨ-r [perf.] / class=rɨχ [imv.] / ma=class=ʁʷˤ [prohib.]. In imperf., with polysemy: 'to go / to come'.
Suppletive paradigm: class=iʔi ~ class=r=uʔ-u-r- [imperf.] / class=ɨx-ɨ-r [perf.] / class=rɨχ [imv.] / ma=class=ʁʷˤ [prohib.]. In imperf., with polysemy: 'to go / to come'.
Note the correspondence qʼˤ - ʔ in the perfective stems between Ixrek class=r=uqʼˤ-u-r- and Mukhad/Luchek class=r=uʔ-u-r- (the Ixrek form is confirmed in both [Dzhamalov & Semedov 2006: 85] and [Ibragimov 1978: 215]). Such a correspondence is irregular; it is hard to explain Ixrek qʼˤ. Perhaps a contamination with the perfective stem y=iqʼ-ɨ-r 'to come' q.v. in Ixrek?
It must be noted that for the Khudig subdialect the following paradigm is quoted in [Shaumyan 1941: 148]: Ha-y- [imperf.] / χ-ina [perf.] / χ-i- [inf.] / yiχ [imv.]. Shaumyan's "ayin" in the imperfective stem probably covers ʢ (i.e. ʢa-y- or ʢaˤ-y-). This form is important for etymological analysis, see notes on Common Aghul.
In [Suleymanov 1993: 54], the following Keren infinitive (scil. imperfective) stem is quoted: ʢʷa- ~ ʕʷa- {гIвас} 'to go'. Probably the form from the Usug subdialect.
Tsirkhe and Khpyuk subdialects: infinitive (scil. imperfective) stem ʢʷa- ~ ʕʷa- {гIвас} 'to go' [Suleymanov 1993: 54].
The main etymological riddle is represented by the imperfective forms: Koshan waˤ- 'to go; to come', Gequn waˤ- 'to go; to come', Proper Aghul we-, but Keren baˤ- 'to go'. At least synchronically, these stems do not contain a prefix, cf. the Proper Aghul negative forms de=we 'you don't go' [Magometov 1970: 194 strophe III], da=we-s 'not to go' [Tarlanov 1994: 148] (with the prefixed negative morpheme dV, not infixed).
The correspondence initial Keren b- : Koshan, Gequn, Proper Aghul w- is irregular, because normally Proto-Aghul *b (< Proto-Lezgian *pː, *mː) yields b in all dialects in the initial position [Suleymanov 1993: 66]. Although, indeed, medial and final Proto-Aghul *b shifts to w in Koshan, Gequn, Tsirkhe (subdialect of Proper Aghul), it is normally retained as b in other subdialects of Proper Aghul as well as in the Keren dialect [Suleymanov 1993: 66; Magometov 1970: 36; NCED: 125].
The external Lezgian etymology of Proto-Aghul imperfective waˤ- 'to go; to come' is not entirely clear. It is proposed in [NCED: 134, 572-573] that Aghul waˤ- (Keren baˤ-) originates from Proto-Lezgian *ʔiqːʷˤä- with the unique case of the development *qːʷˤ > Aghul wˤ (Keren bˤ), which does not seem apt per se.
We assume a more complicated scenario. The Proto-Aghul imperfective stem for 'to go; to come' indeed goes back to the Proto-Lezgian root *ʔiqːʷˤä-, where *qːʷˤ regularly yielded a uvular or laryngeal (ʁʷ, ʕ or ʢ depending on a dialect, see [NCED: 132-133]). This sound is retained in the prefixless imperfective form in the Khudig Koshan, Keren, Fite and Proper Aghul dialects. In other dialects (including the Burshag subdialect of Koshan) the perfective stem was modified with a labial prefix, which supplanted the initial laryngeal/uvular, - as a result, forms of the shapes waˤ- and baˤ- arose in individual dialects. The origin and the phonetic nature of this labial prefix are unclear, but most probably it is related to the well-attested Aghul verbal spatial prefix f= 'near', for which see [Magometov 1970: 158 ff.; Suleymanov 1993: 161]. The semantics 'near' fits the attested meaning of imperfective waˤ- ~ baˤ 'to go; to come' rather well. On the other hand, cf. the prefix w= '?' in the Koshan verb w=aχʷa- 'to speak' (see notes on 'to say').
Such a scenario implies the following archaisms in individual Aghul dialects: the prefixless root is retained in the Khudig (subdialect of Koshan; Ha-y-, i.e. ʢaˤ-y-?), Keren (ʢʷa- ~ ʕʷa-), Fite (ʁʷˤa-) and Proper Aghul (Tsirkhe, Khpyuk ʢʷa- ~ ʕʷa-) imperfective stem, as well as in the prohibitive stem in Burshag (subdialect of Koshan; ma=ʁa ~ ma=ʢe, the loss of pharyngealization and the fluctuation ʁ~ʢ are not clear, however).
The same in the Khanag subdialect: aqːˤ-ˈ [imperf., inf.] / wuš- [perf.] 'to go' [Uslar 1979: 595, 631, 1002] (in [Dirr 1905: 156, 230], quoted as aʁˤ- [imperf., inf.] / uš- [perf.] 'to go' - either actually a form from some Southern Tabasaran subdialect or the beginning of the phonetic process qː > ʁ in Khanag during the 2nd half of the 19th century between Uslar's and Dirr's records).
The same in the Khyuryuk subdialect: aqːˤ-ˈ [imperf., inf.] / ˈuš- [perf.] {аьбкъюв} with polysemy: 'to go / to go away' [Genko 2005: 22].
For the Khiv subdialect two semantically close verbs are documented: aʁˤ-ˈ [imperf., inf.] / ˈuš- [perf.] {аьгъюб} 'to go (backwards)' [Genko 2005: 22] and ʁˤ=aʁ-ˈ [imperf., inf.] / ˈuš- [perf.] {гъягъюб} 'to go (with a certain purpose); to go away' [Genko 2005: 46] (initial ʁˤ= is a spatial prefix).
There is also the plain Khiv stem ʁˤa- 'to go', attested in two morphologically strange forms: imperf. ʁˤa-ra- (< *ʁˤa-ura-?), inf. ʁˤa-b {гъяб} (< *ʁˤa-ub?) [Genko 2005: 46; Khanmagomedov & Shalbuzov 2001: 119]. Apparently this is a rare verb.
Similarly in Literary Tabasaran: ʁˤ=aʁ- [imperf., inf.] / ˈuš- [perf.] {гъябгъюб} 'to go' [Khanmagomedov & Shalbuzov 2001: 119] (for the perfective root cf. ʁ=ˈuš-un-za 'I went' [Khanmagomedov & Shalbuzov 2001: 293 sub uzu]).
The imperfective stem aɢˤ- / aqːˤ- / aʁˤ- is obviously related to Tabasaran ɢˤ- / qːˤ- / ʁˤ- 'to come' q.v.; therefore, it is natural to analyze aQˤ- as a=Qˤ- with the spatial prefix a= (i.e. ʔa=) 'in'. Cf. also the prefixless stem ʁˤa- in Southern Tabasaran (Khiv), if Genko's data are reliable.
Etymologically, the same paradigm in Literary Lezgi: fi- [imperf.] / fe- [perf.] / alˈad [imv.] {фин} 'to go' [Talibov & Gadzhiev 1966: 335; Gadzhiev 1950: 253; Haspelmath 1993: 488, 520; Gyulmagomedov 2004, 2: 315]. Distinct from literary prefixed qːe=qːʷe- {къекъуьн} 'to walk' [Talibov & Gadzhiev 1966: 185; Gadzhiev 1950: 915].
Note the different treatment of Lezgian *ɬʷ in Uslar's dialect (š before old front vowels) and literary language (f). The same development is observed in the Yarki dialect (Kyuri group): Nyutyug šu- 'to go' [Meylanova 1964: 76]. The reflex š is not noted in [NCED: 148 f.].
NCED: 572. Distribution: The situation with Lezgian basic verbs of motion, 'to go' and 'to come', is rather tangled. Almost all the languages possess suppletive paradigms with two etymologically different stems: imperfective and perfective (frequently, a third imperative stem is also present). It is possible to regard this suppletion as a late innovation of areal origin, but there are actually no reasons not to project the pattern of lexical opposition between imperfective and perfective stems onto the Proto-Lezgian level.
The following protoroots are attested with the generic meaning 'to go' in Lezgian languages:
| 'TO GO' | Proto-CA-Udi | Archi | Kryts | Budukh | Tsakhur | Rutul | Aghul | Tabasaran | Lezgi |
| *ʔiqːʷˤä- [NCED: 572] | imperf. | imperf./perf. | imperf. | imperf. | perf. | imperf. | imperf. | ||
| *ʔVʔʷV(r)- [NCED: 1016] | imperf. | imperf. | |||||||
| *ʔišːä- [NCED: 656] | perf. | perf. | |||||||
| *ʔiɬʷe [NCED: 664] | perf. | perf. | perf. | imperf./perf. | |||||
| *ʔa(r)kʼɨ [NCED: 267] | perf. | ||||||||
| ? | *ac- perf. |
| 'TO COME' | Proto-CA-Udi | Archi | Kryts | Budukh | Tsakhur | Rutul | Aghul | Tabasaran | Lezgi |
| *ʔiqːʷˤä- | imperf. | perf. | imperf. | imperf. | imperf. | ||||
| *ʔVʔʷV(r)- | perf. | imperf./perf. | imperf. | ||||||
| *ʔarƛːe- [NCED: 422] | imperf. | perf. | |||||||
| *ʔišːä- | imperf. | imperf. | |||||||
| *ʔiɬʷe | perf. | perf. | perf. | ||||||
| *ʔirqʼär- [NCED: 268] | perf. | ||||||||
| ? | atːˈa- perf. |
It must be noted that in [NCED: 268], the Udi imperfective stems of the verbs 'to go' and 'to come' are connected to Lezgian *ʔirqʼär-, because the authors were not aware of the Udi and Caucasian Albanian forms that retain -ʁ-. Now this etymology can be rejected. The obvious source of the Caucasian Albanian-Udi verbs for 'to go' and 'to come' is Lezgian *ʔiqːʷˤä-. Not a single Udi etymology with Lezgian *qːʷˤ has been proposed by the authors of [NCED], but the development *qːʷˤ > Udi ʁ(ˤ) is system-predictable, see the table in [NCED: 132]; further, Udi ʁ > 0 in certain positions. The main difficulty is the absence of pharyngealization in the Caucasian Albanian and Udi forms, but, as noted in [NCED: 143], the old pharyngealization can sporadically get lost in Udi (exact conditions are not clear).
The Proto-Lezgian imperfective stem can be safely reconstructed as *ʔiqːʷˤä- with polysemy: 'to go / to come'. It retains the original meaning in both outliers (Udi, Archi) and in some Nuclear Lezgian lects. In the rest of Nuclear Lezgian languages, it got superseded with various verbs of movement. It is theoretically possible that the meanings 'to go' and 'to come' were discriminated with the help of spatial prefixes in Proto-Lezgian (cf. the relict prefix in the Caucasian Albanian-Udi imperfective stem for 'to come').
Reconstruction of perfective stems is less obvious. There are two formally equivalent candidates for the Proto-Lezgian perfective stem 'to go': (1) *ʔiɬʷe, which can be reliably reconstructed as the Proto-Nuclear Lezgic perfective 'to go', (2) Proto-Caucasian Albanian-Udi *ac-. The latter root seems, however, etymologically isolated within both Lezgian and North Caucasian, and therefore, we tentatively posit *ʔiɬʷe as the Proto-Lezgian perfective 'to go'.
The best candidate for the Proto-Lezgian perfective stem 'to come' seems to be *ʔVʔʷV(r)-, which retains its basic meaning in Udi and West Lezgic (Rutul, Tsakhur).
Numerous semantic shifts between verbs of movement in individual Lezgian languages require additional investigation.
Reconstruction shape: Correspondences seem regular except for the loss of pharyngealization in Caucasian Albanian-Udi.
Semantics and structure: Primary verbal root, attested with several grades of Ablaut. Reconstructed as the imperfective stem with polysemy: 'to go / to come'.
| Help | ||||||
| |||||||
