COMMENT: Reconstructed for the PEC level. The hypothesis of derivation from *=ixA 'to reap' (Klimov 1971, 227) is tempting, but there are some obstacles: medial *-n- (reconstructed because of the supposed PL *-n- and of PN and Av. initial n-); labialisation of *xw; difference in tenseness. It is probably wiser to think of a secondary contamination of the two roots in some languages (Av.-And. and Darg.).
COMMENT: Reconstructed for the PEC level. The comparison is very probable, although strange phonetically: a lot of irregularities in the development of *ƛ in all subgroups. Maybe this is due to a special development of a rare cluster *-rƛ-? A metathesis similar to the one that occurred in PA, apparently happened also in HU: cf. Hurr. šummǝ 'hand' (see Diakonoff-Starostin 1986, 23).
COMMENT: The EC-WC comparison was proposed by Trubetzkoy. Dissimilatory delabialisation after an initial labial is normal in PWC, as well as the palatalized reflex *Ĺ (since non-palatalized *L is not reconstructed). A minor phonetic problem is, however, the preservation of nasal *m- in PWC (a non-nasal *P- would be expected before the following -r-resonant). It is not excluded, however, that *-r- in the PL form *merƛ̣:ʷ is secondary (duplicating the -r- of the oblique stem *meƛ̣:ʷV-rV- > *merƛ̣:ʷV-rV-, cf. modern obl. stems like Ag. merk:ur-a-, Bursh. mirk:-ri-, Rut. mɨkɨrɨ-, Kryz. mɨkɨr-). In that case *mäƛ̣_wɨ̆ should be reconstructed for PEC, and the PWC correspondence would be wholly regular.
See Trubetzkoy 1922, 241, 243; 1930, 276; Abdokov 1983, 97.
COMMENT: Reconstructed for the PEC level. Tenseness of -x- in Lak. is unknown (attested only in the syllable-final position), thus a reconstruction of both *-ʎ- ( > Lak. x) and *-x- ( > Lak. x:) is possible.
A reconstruction *bĕnʎ_V is also possible; in this case one could also compare HU *pāl- (Hurr. pāl-, Urart. pal- or pāl-) "to pronounce, to tell" (see Diakonoff-Starostin 1986, 17).
COMMENT: Not quite clear is the loss of liquid in PL; otherwise all correspondences are regular. The meaning "oak" in PTs was probably influenced by another EC root, *mħŏq_wĕ "oak-tree, acorn" q.v.
COMMENT: Reconstructed for the PEC level. We must note some interlingual loans: 1) since the Lak and Darg. (Chir.) forms have an identical reduplication, it is probable that one of them borrowed from another (probably Darg. < Lak., since the form is known only in the Chirag dialect of Darg.); 2) Archi has a word mam 'raspberry', usually in the combination gʷac:ilin mam (lit. "mare's nipple"). It is probable that Archi borrowed the Lak. form mamar(i) and reinterpreted the -ar(i) part as the plural suffix, thus building a neologism mam (coinciding phonetically with mam 'nipple, breast' which is of course quite coincidental).
A form like PN *mur-iḳ (*mor-iḳ), reflected in Chech. mürg, was probably the source of Osset. murḳä 'guelderrose', further - Ad. marāḳʷa, Kab. mārḳʷa 'strawberry, blackberry', Ub. mǝ(r)ḳ́a- id., Abaz. maraḳʷa 'mulberry', Balk. marako 'strawberry' etc. This direction of borrowing is probable because in PN the element -iḳ is easily recognized as a suffix.
See Abayev 1973, 141; Shagirov 1, 266 with literature.
COMMENT: For PNC we can reconstruct also the obl. form *mĕrs_A-lV- (reflected in several Lezg. forms, Khin., AAnd. and Lak.). The Laki form lis:u reflects a regular dropping of initial *mV- (lis:u < *mVlis:u, the latter being a metathesized form of *mV(r)s:i-lu).
The WC forms are very likely to belong here, too, although problems are raised by initial *p:- (we should expect *bǝsa); it may be a result of irregular dissimilative development within the PWC compound **bǝ̃la-bǝsa.
COMMENT: Reconstructed for the PEC level. Tense *ć̣_- reconstructed because of the Nakh reflex (after a weak vowel *ć̣ would yield PN *-ss-). PN *merc̣ is obviously a transformation of *mec̣rV- (analogously from the oblique base to direct - a rather frequent process). Since there are no Av.-And.-Tsez. reflexes and the root is polysyllabic, the vocalic reconstruction is not very certain.
COMMENT: Reconstructed for the PEC level. The u(ū) vowel in PN is probably due to progressive labialisation of the initial m-. On a possible trace in Lezgian (Udi) see under *wǝrqV "sheep's tail".
COMMENT: A Tsez.-Lak. isogloss. Perhaps, one could compare also Av. maṭú 'mirror' (borrowed into many EC languages: And., Akhv. Cham., Tind., Inkh. maṭu; cf. also dialectal Av. (Chad.) muṭú id. and Gunz. muṭu).
COMMENT: Reconstructed for the PEC level. The Nakh, Darg. and Tab. forms are hard to separate, although we have to assume different as-/dissimilations to unite them (in PD obviously *wana- < *mana-, because genuine w- is very rare in PD; and either PN *mɦal- through dissimilation < *mɦan-, or PD, PL *manV- through assimilation < *malV-). Since interaction of this kind between two neighbouring resonants is very frequent in Caucasian languages, we think that the etymology is rather probable.