From "Chan Grammatical Analysis" to "Laz-Megrelian Grammar" – towards the History of the Interrelationship of Laz and Megrelian

Laz-Megrelian study has a long history. Still, many issues of its structure and history require clarification and in-depth research. Just to this problem is dedicated the project fundamental work "Laz-Megrelian grammar. I. Morphology "which demonstrated the correlation, similarities and differences between Laz and Megrelian. The research outcomes confirmed a substantiated and stable nature of the postulate: system of sound correspondences is the basis for the correlation between Laz and Megrelian as linguistic entities. Linguistically, the limits between the Kartvelian languages and dialects are determined by regular, systematic and logical sound correspondences which were revealed among the members of this language group. Vowels and consonants demonstrate common picture in Laz and Megrelian, the language (phonological) system is similar. Therefore, Megrelian-Laz should be deemed to be two dialects of one language. Similar qualification is supported by the morphological analysis carried out by the project participants. The work deals with the correlation between the Megrelian and Laz as well as their linguistic status.


Introduction
The project -"Laz and Megrelian comparative grammar (I. Morphology) with the index of common roots and stems (N11.07)" , which was funded by Shota Rustaveli National Science Foundation, scrupulously, comprehensively analyzed Laz and Megrelian systemic grammar issues, summed up what had been investigated, a number of key issues are studied in a new way. Based on fundamental research outcomes as well as the achievements of modern Kartvelological studies the work analyzed problematic issues of the Laz-Megrelian grammar.

Methodology
The issue has been processed from synchronic and diachronic (historical) standpoint. Scientific research used descriptive, historical and comparative methods, which essentially were defined by the tasks set.
Laz and Megrelian subdialectal properties have been considered.
Comparative analysis of the Laz and Megrelian data demonstrated the quality and outcomes of systemic changes.
The study is based on new field materials, which were obtained by the authors and have not been in scientific circulation up to date.

Results
The Megrelian-Laz morphology is characterized by one type declension: a) The number of simple cases of both linguistic units is basically similar: always loses the initial vowel marker due to the influence of a stem vowel. An originally Nominative case marker -e that is merged with a stem can similarly impact on a vowel case marker in Laz-Megrelian. This fact demonstrates once again that -e belongs to a stem and does not occur in the declension process.
The fact that initially -e really was a Nominative case marker (along with -i) is reinforced by the reality that Laz plural marker of plural nominals is always represented by -ep-e form, unlike Megrelian, which uses -ep-i like in Georgian (= Geo. -eb-i): koC-ef-e, Megr. koC-ef-i 'men'. The other features are also observable, in particular, in Megrelian the nominals with -a vowel-ending are elided due to the influence of a plural marker -ep, but in Laz, on the contrary, like in the declension process, a grammar tools (resp. a case marker with a vowel at the beginning) undergoes changes -Megr. dixa / dix-ep-i 'lands', Laz. dixa / dixa-p-e 'lands', but in the case of nominals with o, e, u vowel-endings the system undergoes correction, even in Megrelian -e of plural marker is lost.
The Laz and Megrelian reveal similarity in formation of plural of da-l-ep-e/i 'sisters' type which is known in the professional literature: an initial root -l sonor, which is elided in singular, is maintained in plural forms, like in word derivation process; cf. Megr.-Laz. Ja 'tree' (= Old Geo. Zel-i) -pl. Jal-ep-i/e 'trees', o-Jal-eS-i 'for beam', Jal-ona 'alley' ... c) in declension of a determinatum and determinandum in both Zan dialects, there is no difference between the stems with vowel and consonant endings: neither in Laz nor in Megrelian during the declension of both nominals a determinandum does not change a case, accordingly, a number is not changed, as well. This Laz-Megrelian phenomenon is so obvious that even G. Rosen notes that the determinatum and determinandum is declined like a compound word: didi koCi 'big man", didi koCi-k 'big man', didi koCi-s 'big man', didi koC-iS 'of a big man'... Although the plural nominals with -ep-e and -ep-i markers seem to be homogenous in Laz and Megrelian, the relict traces of plural with n-and t-markers at the demonstrative pronouns in Megrelian as well as in Laz seem to be especially noteworthy phenomenon; cf. Laz ha-n-i 'these', hi-n-i 'those' (like in Georgian) and the mu-n-ep-i 'those' in Megrelian, while mu-t-ep-i 'those' in Laz. A singled out -n-in Megrelian and Laz and -t-in Laz are believed to be formants of plural with -n-and -t-because a suffix -t-a is evident in certain Megrelian syntagms.
The forms of a superlative degree are more numerous in Megrelian, diversion is usually productive. In Laz their number is less, but affixation is similar -u--aS confix, a morphological and phonetic correspondence of a Georgian u--es prefix and suffix, is maintained in both subsystems: Laz-Megr.
u-mkil-aS-i/u-kil-aS-i 'the youngest', u-did-iS-i/u-did-aS-i 'the biggest/largest', u-nC-aS-i//u-C-aS-i 'the eldest'... At that time we have the only exception -u-mCan-e 'elder/senior', which is attested in the superlative degree form only once, cf. Svan. meC-i 'old man'.
In Zan dialects a formation of an equality degree is revealed, as well: Megr. ma-Sxu-a // ma-Sxv-a 'as thick as…', Laz. ma-peJan-a "id". The formation of a comparative degree with -ora suffix (mo-//do-)myit-ora 'reddish', which seems formally different from /mo--o/ : /mo--e/ ((mo-Sav-o : mo-uC-e 'blackish') confixal derivation attested in Georgian and Megrelian, seems to be notable. In the terms of origin, our attention is focused on Svan -ara (sgel-ara 'thickish') suffixal derivation: an exact correspondence of a Laz suffix is revealed: Laz -ora: Svan -ara. It is noteworthy that in Laz an initial position is similar to Georgian-Megrelian -a preverb is used, the final affix coincides with Svan one and reflects *-ara morpheme of the Common Kartvelian radical language.
Morphologically, a verb can be either single-personal or bi-personal in Laz and Megrelian. In a single-personal verb only a subject person is expressed, in bi-personal -an object and subject, in tri-personal verb, like in a bi-personal one, only one of objects (case-unchanging). We emphasize, that according to morphological parameters a Laz-Megelian verb is bi-personal. From this standpoint Laz and Megrelian reveal similarity. Both dialects manifest difference in relation with Old Georgian and Svan where a verb expresses the third person, as well, even if it expresses a case-changing plural object: in Old Georgian --n, in Svan --al.
The Inversive process of relative and static verbs is clearly shown in Laz-Megrelian, since inversion is associated with a case-unchanged object. Accordingly, an issue of morpho-syntactic function of a Dative case in Megrelian and Laz arose. In Megrelian the function of Dative is similar of that in Georgian: it is a case for a case-unchanging as well as case-changing object, also a real subject of inversive transitive verbs in III series and RS of relative-static verbs are in Dative case, i.e. Dative case position in Megrelian is stable. The same cannot be said about Laz. In Laz only case-unchanging object person and real subject of inversive relative-intransitive verbs are in Dative case. It appeared that in transitive verbs inversion is less productive -another system to form the Perfect tenses has been established. But despite the fact that in Laz the constructions of objective order of relative-static verbs are often violated, the morphological structure of a verb is unchangeable. Thus, in Megrelian and Laz a common inversive model is formally used. In particular, a real subject is represented by personal markers of m-articulation zone, and a real object -by the third subject personal marker.
In Laz-Megrelian the plurality of a real subject in a verb is expressed by -an suffix in the Present and Future scrieves, -es suffix -in the Past scrieves ; -n is expected in Subjunctives: uGut-a-n (= Megr. uGud-a-n) 'they'd have', aSquin-a-n (= Megr. aSqurin-a-n) 'they'd be scared'...
A structure of a preverb in Megrelian and Laz is basically similar. In both a simple preverb consists of a consonant and vowel. But an exception, i.e. a preverb is expressed only by a vowel is also attested. The simple preverbs e-, a-, o-are similar, as well. A compound preverb is of following structure: consonant + vowel + consonant + vowel or vowel + consonant + vowel. All adverbs contain a vowel, Laz n-adverb is exception. In Laz-Megrelian a system of preverbs is represented by two subsystems. On the one hand, there are simple preverbs denoting a concrete direction: upwards e-; downwards -ge-/gi-; thither -me-/mi-; hithermo-/mu-; around -go-; do-preverb is an alternative of preverb ge-/gi-, but also it has other functions, as well (intensity, repetition ...): do-lu-n 'is falling (down)'; do-Tolums (= Megr. do-Toluns) 'will cut/slice up' ... On the other hand, it appeared, that like in Laz, the same elements derive compound preverbs in Megrelian. However Megrelian lacks ja-, jo-, xo-elements, however, it has more to-and no-segments. Among Laz-Megrelian simple preverbs a prefix o-preposition is a special case, which cannot express a direction and mostly derives a perfect aspect forms: ykomu 'eating' -o-ykomu 'ate', yofu 'catching' -o-yofu 'caught' ... A secondary character of aspect formation with a preverb is approved by that group of verbs, which, despite the adding some preverbs, express an imperfect aspect: dgun 'is standing' -ge-dgun 'is standing'; xen "is sitting' -ge-xen 'is sitting', me-Sa-xen 'is sitting inside' ...
With regards to the verb conjugation the work presents the following: conjugation changes a verb according to scrieves. And just such changing is deemed to be the basis for a category of conjugation, i.e. in a verb morphology a scrieve is the same as a declension in a nominal morphology. Obviously, any verb potentially occurs in one of existing forms. The unity of these different forms makes a basis for conjugation category. Thus, the conjugation is a general term denoting a category, a scrieve is a term denoting a concrete verbal form (i.e. Perfect scrieve is the same morphological data, as an Ergative case; only, they are the members of different systems).
The comparison of homogeneous verbal forms attributes a conjugation to a category. Grammatical (morphological) naming of a member of oppositional relation of Megr.-Laz. kokiminu 'did/made' type is a scrieve of perfect form and its naming cannot be a category simultaneously. Analogously: a grammatical (morphological) name of a member of an opposition relation of Megr.-Laz. oxorik 'house' is an Ergative case form and its naming cannot be a category, as well. If it concerns a category, it should be discussed about a category of conjugation and not of a scrieve. In other words, 374 From "Chan Grammatical Analysis" to "Laz-Megrelian Grammar" -towards the History of the Interrelationship of Laz and Megrelian any case form is a member of an opposition relation and, therefore, it makes the basis for considering a declension to be a category. Additionally: a form of a concrete scrieve is a member of opposition relation of other system and therefore, is the basis for considering the conjugation to be a category. Each scrieve form expresses the grammatical meaning. All scrieves have a form. This form is created through a certain combination of morphological elements. The fact is that the present scrieve is different from the aorist and perfect one, continuous -from future, future -from subjunctive, etc. A structure of verbal units should have the function of a morphological marker. The fact is that this is an orderly system. Like a nominal has a category of declension in parallel with other categories -morphological category (of number) and semantics (animate, inanimate ...), also a verb has a category of conjugation in parallel with other morphological category (of number, voice, version, contact, aspect ...). As regards a conjugation category it will be a form of any scrieve, according to a person category -a form of any person, according to a number category -one of number forms, according to a subject's activity-passivity category -a form of active or passive, according to the category of possession and purpose -of subjective, objective and neutral version, etc.
The classification principle is the following: declension is perceived as a unit of a formal oppositional relation in a declension category. A scrieve is a unit of a formal oppositional relation in a conjugation category. The formal relation implies that this relation has function (i.e. when we speak about a form, we consider not an abstract form but a content of a form). Otherwise, speaking about a category is groundless. Like in Georgian, in Laz a system of scrieves is presented not according to formation principle, but according to a functional one: "Formation of scrieves implies the formation of stems and, therefore, in scrieves system any scrieve should occupy that place which belongs to it in terms of formation" (Arnold Chikobava).
The work discusses the issue of Laz Present in a new way, in particular, the forms of nimers "is taking away", otkomers "is throwing" type aren't historically considered to be stem-changeable verbs as an initial stem is the same for the forms of I and II series, so we consider a nimers form to be a secondary form, an initial seems a *niG-mers variant, which revealed a root morphemes (cf .: Aoristi: niG-u); G + m and any other consonant group (C + m) form rising openness complexes in Present which in Laz and Megrelian (as well as in Proto-Zan) were overcome by various means, among which one was superration of non-canonical order due to the influence of a next sonor caused by the falling/loosing of a root vowel, like Geo. blikvi 'fist' → Megr. likvi and Laz. *qoSematkobmers > qoSematkomers 'will hide smth in/into', *niGmers → nimers 'is taking/takes (away)'. Accordingly, Aorist presents a root consonant: qomeSatkobu 'hide smth in/into', niGu 'took away', but Present loses it due to a Laz (Zan) phonotactics.
Laz separates a Megrelian-like scrieve of Future with a preverb as a system, which is based on Present and similar to that of the Georgian-Megrelian Future. In the Laz language the Present with either a preverb or qo-/va-particle is a Future form and occupies a proper place in the conjugation system, which is already presented. A paradigm of Future scrieve of a Zan verb yar-'writing' is an illustrative example, especially when such formation is frequently fixed in published Laz texts: Megrelian p-yar-un-q 'I'm writing' do-p-yar -un-q 'I'll write' yar-un-q 'You're writing' do-yar-ur-q 'You'll write' yar-un-s 'He/She is writing' Laz do-yar-un-s 'He/She'll write' The view sustained in this paper is supported by a scrieve paradigm of Future Aorist of yar-'writing' verb, when the system similarity is beyond doubt: 'You were writing' do-yar-up/um-ti 'You'd be writing' yar-up/um-t-u 'He/She was writing' do-yar-up/um-t-u 'He/She'd be writing' In both languages -in Megrelian and Laz lthe Future Subjunctive is a Present Subjunctive form complicated with a preverb (Megr. do-p-yar-un-d-a, Laz. do-p-yar-up/um-t-a 'I'd write'). It is followed by a Future Conditional, that seems to be Present Conditional complicated with a preverb (Megr. do-p-yar-un-d-i-ko(n), Laz. do-p-yar-un-t-i-ko(n) 'If I'd write').
Thus, in Laz a scrieve system of I Series is similar to that in Megrelian, which implies its systemic similarity to Georgian. However, unlike Georgian, Megrelian-Laz has more specific forms conditional scrieves.
A system of scrieves which is illustrated in two verbal roots yar-'writing' and Gur-'the death', is as follows: The comparison with the traditional system will demonstrate that the difference is obvious, but it is based on appropriate arguments. Actually, so called the Future derived from Subjunctive has been removed from Laz traditional system of scrieves, but its place was occupied by the Future formed analogous to that of Megrelian and which is derived through the combination of a preverb of neutral semantics and present form. In this way Future forms have been restored.
Those forms have been removed from a system of scrieves, the components of which were: unon / untu / onteren. We considered that they are the elements expressing the semantics of desire/wish with which they invest corresponding verbal forms, which independently express the grammatical contents of any scrieve.
The forms of zumatu / zumatukon type have been removed from a system of scrieves, as the forms of independent scrieve, because we suppose that a case-marker seems to be simplified in zumatu, yaratu... zumatukon, yaratukon... derived forms: zumatu < zumaptu... zumatukon < zumaptukon. Such an explanation cannot be blocked by the fact that in zum-um-s form -u-(and not -a-) occurs. A case-marker changing is quite common process not only in Zan, but generally, in the Kartvelian languages. Therefore, zum-a-tu is considered to be a parallel version of zum-um-tu; analogously: zum-a-tu-kon < zum-ap-tu-kon... In a word, the forms of bzumati, bzumatikon type which are the result of phonological simplification, underwent changing: -ap > -a. Consequently, on this ground a bzumati form should be logically considered not a separate scrieve, but as a variant of Imperfect, analogously, bzumatikon -phonologically simplified variant of Imperfect Conditional. A system of Laz scrieves is similar to that of Megrelian not only in the number: Laz, as well as Megrelian has 19 scrieves. The eight out of nineteen scrives are characteristic of Laz-Megrelian (including 4 scrieves -so called the forms of IV Series are characteristic of some Georgian dialects). The rest are common with Georgian.
As for the morphological category of voice from andpoint of structural (formal) analysis of a verb, we consider that Zan has only two voices: passive and active. For the verbs of a middle voice, T. Uturgaidze's view should be taken into account: these verbs (qris 'is blowing', wuxs 'is sad/worries', duGs 'is boiling', dgas 'is standing' , wevs 'is lying' ...) became medium verbs after the formation of voice category. The verbs of middle voice are still separated, only because they form a group of verbs, which can be called unvoiced verbs, because they have no voice. This was partly determined by their semantics, and largely by their static character. These verbs have not developed a voice and, only after this it became possible to distinguish them as a separate group, since it is obvious that in middle-active verbs a direct object is lost and owing to this the action, which moves from a subject to an object, becomes a subject-receiver and the impression is that we deal with other type of a verb, while the difference is made by losing only one unit, cf. Old Geo.
h-Kiv-i-s _ h-Kiv-a (!), but Mod. Geo. Kiv-i-s _ i-Kivl-a 'is crowing --crowed'. In Megrelian-Laz the Laz verbs of oxorJa-q-i-xoron-s / xoron-ap-s 'a woman is dancing' with iprefix and case-marker (am / -ap, -um / -up ...) type are considered to be dynamic (active) verbs in both language subsystems, and due to this they adapt to a subject in Ergative case, otherwise static verbs never occur in Ergative case. Static-dynamic character of the verbs without a case-marker determines the absence or presence of the aorist of tenses and moods, in particular, the verbs with a case the History of the Interrelationship of Laz and Megrelian marker have aorist, the static verbs without a case-marker and with -e/-u suffix -never have one. This is usually a stable rule in Laz-Megrelian. The Laz-Megrelian material showed the tendency for activation (turning into dynamic) of static verbs. This tendency is clearly featured in Laz material. As it is known, Laz has no -un case-marker, it is characteristic only of Megrelian and occurs when in Laz -um is attested (in Khop. -up). The latter is an exact phonetic-functional correspondence of Georgian case-marker -am (→//-av). But -un is also attested as a case-marker in Laz, but its origin is different and is the result just of the activation of a stem, which is supported by the existence of parallel forms in Laz. An example of a static verb with a -u-suffix futx-u-n ' is jumping/leaping, flying' -kvinCi futx-u-n 'A bird is jumping/leaping'. As it was expected, a subject of a static verb is in Nominative case, but by adding a personal marker -s of a S 3 , which results the activation of a grammatical content of a verb, makes a verb unfamiliar for Laz (a verb with -un case marker), that becomes dynamic and in Present a subject occurs in Ergative, like all other transitive verbs: kvinCi-q futx-u-n-s 'A bird is scared'.
Conjunction should be analyzed in terms of syntax, but in general, it can be viewed morphologically, as well (Shanidze). At first, it refers to the structure and origin of interjections. Virtually, the history of studying the Zan conjunctions started N. Marr's "Chan (Laz) Grammar", in which the scholar discusses several conjunctions, their place in a sentence [9,71]. The work demonstrates the correlations between Georgian and Laz-Megrelian grouping, separating, opposing and subordinating conjunctions.
In Laz-Megrelian the causative semantics is expressed by a morpheme -ap, which coincides with a derivative formant of a present stem, i.e. a case-marker. In a causative form, along with -ap suffix a prefix o-is also confirmed which is a morpho-phonetic correspondence of Georgian -a. They have a common grammatical function -to express not only causative. In this regard, in Megelian-Laz the picture is same as in Georgian. Unlike in Laz, in Megrelian -ap suffix is followed by -u. Megrelian causative verb has -apu-ending in I series. In Perfect in a relevant phonological surrounding, in Megrelian there occur distanced double -ap: uyar-ap-u-ap-u 'Smb made smb write'. In Khop subdialect of Laz there occurs -ap-up combination in the present scrieves, in Vits-Arkab -ap-am combination. In Megrelian-Laz the present should be initial for those causative forms in which root form an/-am/-ap thematic suffix is attested. Just this suffix is considered to be a marker of causative.
Infinitive and participle are derived similarly in Megrelian and Laz. N. Marr argued about "an obvious difference" between basic derivative formants of masdar (in Megrelian -ua -in Laz o-u confixes) in order to consider Laz to be a separate language. But it turned out that a masdar with a Megrelian-like -ua suffix is also attested in Chkhal sub-dialect of Laz. Chkhal regularly reveals -ua ending. This fact brings a Chkhal sub-dialect of Laz closer to Megrelian:

Discussion
Arnold Chikobava's work "Chan grammatical analysis" [1] is the first work in which Laz-Megrelian-Georgian language systems are compared. This work as well as "Laz-Megrelian-Georgian Comparative Dictionary" [2] laid the foundation for studying the Kartvelian languages, which is directed to study Laz-Megrelian as the Zan language dialects. Since then the work has been carried out at the TSU Arnold Chikobava Institute of Linguistics, while simultaneously, basing on G. Deeters, one part of researchers have been studying Laz and Megrelian as independent languages. This work was particularly fruitful in the 1960-ies: Th. Gamkrelidze and G. Machavariani's work "System of sonants and ablaut in Georgian" [3] was published. Since then many fundamental works, articles have been created and large noteworthy material have been accumulated. The present level and direction of linguistic research required comprehensive and in-depth comparative linguistic analysis of Laz and Megrelian, which would be an important step to compile comparative grammar of the Kartvelian languages. This was essential to interpret newly a number of morphological issues, which were considered as a "watershed" between Laz and Megrelian as independent languages.
In particular, G. Kartozia considered it essential to investigate Laz and Megrelian phonetic rules as well as morphological and syntactic properties. His monograph "Laz language and its place in the system of Kartvelian languages" [4] deals with just the systemic research. According to the scholar, "Zan dialects functioned as one language until morphological, lexical differences were few. In spite of phonological system changes, increasing of morphological and lexical differences caused their fragmentation into two languages" (ibid, 21-22). As the scholar notes, phonetic fragmentation basis of Zan into two dialects is considered some Laz phonetic properties, which differ Khop-Chkhal dialect from Vits-Arkab and Atin-Ardashen, but it is close to Megrelian. As the scholar supposes after the fragmentation of Zan, Khop-Chkhal sub-dialect was separated from Megrelian dialect ("i-dialect"), and "sheltered" Laz ("u-dialect "), due to which it assimilated a number of morphological, syntactical and lexical innovations (ibid, 23). Essentially the same view is given in the monograph "Linguistic analysis of Megrelian" [5], which is an attempt to show those grammatical and vocabulary differences, which indicate the formation of Megrelian from Laz as the independent language.
"Laz-Megrelian grammar. I. Morphology" [6] prepared and published at Arnold Chikobava Institute of Linguistics compares Laz and Megrelian morphological systems: to what extent these systems differ from each other, whether these differences can be or not the basis to qualify these languages as independent ones. The monograph analyzes the Laz-Megrelian basic vocabulary that is attached to the monograph as an index of common roots and stems and includes only those lexical units which are common for Laz and Megrelian.
The monograph analyzes Laz as well as Megrelian data. All analytical levels reveals those common and distinctive properties which were characteristic of Laz and Megrelian. Basing on numerous linguistic materials a systematic research reveals those trends which have been outlined in the last few decades in terms of the correlation between Laz and Megrelian, viz: a) in morphology the grammatically or semantically valuable groups have been selected out for each part of speech; peculiarities of form-/word-building have been described; b) morphological classification basis of verbs have been specially processed; verb categories have been indexed: tense system, mood deverbative nominals (masdar and participle); c) form-unchanging parts of speech have been processed and compared on the basis of Laz and Megrelian data; d) index of common Megrelian and Laz roots and stems (actually a comparative dictionary) has been created.
And all these have been processed to show common and difference, the parallels with Georgian have been drawn.
Laz-Megrelian (i.e. Zan) is included in the South-Caucasian (Iberian) group of the Iberian-Caucasian languages, which is called the Kartvelian languages. The professional literature uses other (except Zan and Laz-Megrelian) term to denote this Kartvelian language -Colchian language (A. Shanidze, K. Danelia, Z. Chumburidze ...). The latter underlines language and tribal unity of Megrelians and Lazes since the Colchis era.
The kindred languages of Megrelian-Laz are Georgian and Svan. Formal and functional correspondences at all lingual hierarchical levels turn Laz-Megrelian into the kindred languages of the Kartvelian ones. It makes possible to hypothesize the origin of these languages from a common language and to postulate one proto-language (Common Kartvelian).
From the kinship standpoint, Gerhard Deeters attempted to depict schematically the correlation of the Kartvelian languages and dialects: Truthfulness of this scheme is reflected in its Svan part that implies the formation and separation of Svan (resp. dialect) from a proto-language as an independent linguistic unit, but the Georgian-Zan component (unity?) of the same scheme does not reflect the prehistoric picture, as occurrence the History of the Interrelationship of Laz and Megrelian of sound correspondences (in sibilants as well as in vowels) supports the existence of at least three independent linguistic units (subsystems) in Common Kartvelian. These languages are: Georgian, Zan and Svan. It can be hypothesized that in the prehistoric era there existed the fourth Kartvelian linguistic data. This assumption is supported by those lexical units which occur in the south (Meskhian-Javakhian) dialects of Georgian, but, because of it synonymic character it is considered to be originated from another dialectal group (fourth Kartvelian language). Such word seems a Javakhian zaGar-"gun-dog", which, considering the phoneme relations among the Kartvelian languages, would be characterized as a language of a-sounding, like Georgian and Svan (Geo. ZaGl-, Svan jaGv-"dog"). An initial z is close to Svan spirantization (cf. J → j in Svan and Z → z in the Kartvelian dialect), but the whistling sounding is of Georgian-like. As Zan ( Megrelian-Zan) properties, in alleged dialect (language?) the non-root-syncope property (full sounding ZaGar-) is of Zan-like. A phonetic transformation l → r at the end of the root is of Zan-like, as well (ZaGar -/JoGor-).
The similar features is obvious in another possible lexical unit, like a Geo. axor-"stable" is considered. Even in this case a Georgian-like vocalism, an initial a (axor-/saxl-) is preserved at the beginning of a word, but like in Megrelian-Laz s in anlaut (resp. S) obviously is lost and Zan-like non-syncope feature is again proven, like l → r transformation (cf. saxl-/oxor/axor-). In view of aforesaid, if this assumption is true, two Common Kartvelian archetypes should be reconstructed -*ZaGal-"dog" and *saxol-"house".
In the future, similar material abundance will enable us to discuss the existence of another whistling type language (fourth language-tribal unit) in the Kartvelian linguistics, which should have been of Georgian-like according to other parameters, since the signs of Z → z spirantization-desaffricatization is observable in Old Georgian, as well: zroxa/Zroxa, zaxil-i/Zaxil-i and so forth.
Linguistically, the limits between the Kartvelian languages and dialects are determined by the existence of regular, natural and systematic sound correspondences among the members of this language group. Transformation of the initial (Common Kartvelian) phonological system into three subsystems implies the origin of three language units (firstly of dialect, further of a language) and realization/death of originating language (proto-Kartvelian) in proto-Kartvelian dialects. Denying the Georgian language parameters on the basis of other language (Avar) data is groundless in this case. It is high time to name the linguist factor as linguistics: disclosure of the sound correspondences is a reliable classifier to qualify a language and dialect of the Kartvelian languages and dialects, just it defines the boundaries between Georgian, Zan and Svan, because on the following ground of decomposition of Common Kartvelian proto-language any member of this three-member system was an independent linguistic unit (on the phonology, morphology level).
As the Dagestanian (Avar) language data, in particular, the cooperation between the Tokhur and Chadakol subdialects of the Antsukh dialect where sibilant vowels of these two subdialects reveal the same correspondences as Georgian and Svan-Zan (Th. Gamkrelidze, G. Machavariani, G. Kartozia...). Although, in the named languages there are considerable sibilant oppositions whistling : hushing but sibilant relation of the Kartvelian languages is not based only on whistling/hushing opposition. In this case, along with whistling : hushing correspondence a key role is played by the hushing : hushing + back lingual stop plosive (Geo. Svidi : Zan SqviTi : Svan iSgvid) correlation, since the whistling → hushing is observable not only in the Avar sub-dialects, but in Old Georgian, as well; cf .: frcxili /frCxili, sxami / Sxami and many others, however, not a single case of inter-substitution of J/Jg, C/Cq, y/yk, S/Sq isn't attested either in Old Georgian or other Iberian-Caucasian languages. Generally, determining a language and dialect in the Dagestanian and Nakh systems the phonemic correlations should be taken into account, but the key (decisive) role is still played by morphology (structure). We'll example the Archib language, the phonological system of which is attributed to Avar-Andi-Dido (because of laterals), but from the morphological structure standpoint it is included in the Lezgian subgroup of Dagestanian languages.
The monograph puts the issue of the correlation between the Kartvelian languages in a new way, in particular, it should be expressed as follows: Thus, initially on the previous differentiation level, in the Common Kartvelian proto-language there were at least three dialects -Georgian, Zan, Svan, of which these dialects were formulated as independent (historically fixed) languages on the ground of basic morphological, lexical differences (as well as of phonological transformation).
Transformation of the initial (Common Kartvelian) phonological system into three subsystems implies the origin of three language units (firstly of dialect, further of a language) and realization/death of originating language (proto-Kartvelian) in proto-Kartvelian dialects. Since Laz and Megrelian (and Georgian-Svan) languages reveal the similar phonetic correspondences, they are two dialects of one language, i.e. the origin of Colchian (Zan) from Common Kartvelian proto-language as an independent language has been contributed by sibilant shifting (whistling → hushing, hushing → hushing + back lingual stop-plosive) as well as the transformation of Common Kartvelian vocal system in Zan, what is properly called back shifting of articulation (Arn. Chikobava), Zan shifting of the same vowels [5]. Megrelian and Laz vowels and consonants demonstrate similar picture, the language (phonological) system here is one.
As for morphology: It should be noted that in some cases the difference between Megrelian and Laz is debated, but proper Laz is not always monolithic, namely: a) the third group of tenses and mood in Khop is formed my means of the -ere, -ere-t-i, ere-t-a suffixation, while the tenses and mood is formed by means of auxiliary verb do-r-en, do-r-t-u-n, do-r-t-a-s in Atin, Vits and Arkab; b) in Megrelian a masdar is derived by means of a basic formant -u-a, which is also the only one in Chkhal of Laz. Masdar with o_u formant is natural for Megrelian, as well, while Laz substantized masdars demonstrate a trace of -v-a (← -u-a) masdar, cf. Laz. tax-v-a "corn picking, October" (Megr. tax-u-a), ‡in-v-a/Kin-v-a "winter, cold weather" (Megr. ‡in-u-a "freeze"), wil-v-a "November, picking" (Megr. wil-u-a "picking"); c) one of the distinctive factors იending of aorist in Laz introduced by N. Marr couldn't create a dissonance with Megrelian, since along with the aorist emarker, in non-causative verbs, ი-suffix is an only suffix in Megrelian. Moreover, like N. Marr, Arn. Chikobava noted: "Megrelian uses two suffixes in aorist: -e and -i which separate Megrelian from Chan, which has only one formant -i" [1,[135][136]. But in G. Dumézil's Arkab texts there often occur cases when in formation of an aorist in the verbs with i-prefix, the forms with e-ending are observed in Laz as well as Megrelian: b-i-Sinax-e "I kept", ele-b-i-svar-e "I whittled smth for myself", b-i-qCan-e "I turned white", b-i-Suv-e "I got wet", b-i-GurZul-e "I ate my fill/was sated" ...
A peculiar formation of Future was considered a distinctive morphologic marker of Megrelian and Laz. One of the novelties, which is presented in the work, is just the fact that according to formation rule no difference between Laz and Megrelian (as well as Georgian) is evident, if basing the existing linguistic facts it will be relevantly qualified (the discussion and arguments on this issue a reader will see in the work).
At the same time, I believe that in terms of the dialectal differentiation the current Zan picture reflects the ancient state, when the difference between Megrelian and Laz didn't make the linguistic opposition, but actually other dialectal composition of Zan (Colchian) language (which together with Georgian-Apkhazian-Svan languages covered the vast territory of the Black Sea coastline region -from Sinop-Samsun-Trabzon up to the river Psou) was presented.
As linguistic data reveal, at the next stage of its formation as independent language, Zan (i.e. Colchian) was differentiated into three main dialects. These are: 1. Atin-Vits-Arkab, 2. transition subsystem of Khop type and 3. Megrelian. The first (Atin-Vits-Arkab) dialect covered the present Khopa, the second (Khop type dialect), which apparently is of mixed type, covered the administrative territory of Guria-Samegrelo regions, and the third dialect -Megrelian of Zan (Colchian) language covered the territory from Tskhenistsqali-Rioni-Pichori up to the river Psou.
The fact that Khop type dialect (and not Megrelian) was widespread on the present Guria-Ajaria area is supported by the occurrence of Khop (and not Megrelian) properties in toponymy. For example, in Guria there are attested toponymic Cholchisms (Zanisms) su-xCe and toba-xCe (toba-xCa), in which -xCe/-xCa seems to be the correspondence of Old Geo. ƒce (mƒce) root and is equal to Laz qCe and Megrelian rCe/Ce "white" adjective. Proper Proto-Zanian *ƒCe is the source for Megrelian rCe/Ce via qCe (qC → rC) but from same qCe Khop developed the spirantized xCe allomorph. Just this xCe "white" is attested in Guria-Imereti in su-xCe-form. It is logical , that a Megrelian rCe-cannot be phonetically derived from xCe-(x → r transformation in any combination doesn't occur in Megrelian). Basing on Megrelian phonotactics if Megrelian would have xCe the necessity of its superation wouldn't have arisen. That is why we believe that in ancient times a historical Khop-like dialect was spread on present Ajaria-Guria territory, the traces of which were preserved only in toponymy. Later, due to the loss of direct contact with Megrelian, transition dialect (of Khop type) was restricted on Guria and Ajaria territory -it partially loses Megrelian-like properties and gradually assimilates to Atin-Vits-Arkab subdialect, in other words, it becomes Laz.
So-called historic Khop was a transitive type mixed dialect of Zan (Colchian) language and just it had direct contact with the ancient Armenian, which is reflected in lexical Zanisms in Grabar (Old Armenian).
Modern Zan (Megrelian-Laz) language distinguishes two major dialects -Laz (Chan) and Megrelian. In its turn, Laz is divided into the following subdialects: Atin, Vits, Arkab and Khop. Atin included Bulef-Artashen and Chamli-Hemshin speeches, Chkhal is a separate speech in Khop. In our opinion, Vits and Arkab are independent subdialects and they aren't considered as units of Vits and Arkab subdialects. Arkab is characterized by some specific features, which distinguish it from Vits: a) formation of reduced -don form of -doren auxiliary verb of Present Perfect. In particular, regularly in Present Perfect, which generally is based on Aorist stem in Zan, unlike Vits, zum-es-don, Zir-es-don, yar-es-don allomrphs occur instead of zum-es-doren, Zir-es-doren, yar-es-doren full forms; b) formation of Present by means of -e ending in Arkab, which is quite common, cf .: me-b-a-g-e "I guess", oko-b-i-xv-e "I separated", do-b-a-rC-e "I'm running", Toli o-b-u-dum-e "I'm closing the eyes". In this cases, a Present stem is simplified and a final r disappeared without a trace, i.e. in the first three verbs there should have been r suffix characteristic of static verbs: me-b-a-g-e-r, oko-b-i-xv-e-r, do-b-a-rs-e-r; an active voice verb o-b-u-dum-e loses a final r (resp. mer), as in other cases: dovuwume ← dovuwumer, qiSime ← qiSimer, dobdume ← dobdumer; c) as aforesaid, a common feature of Arkab subdialect and Megrelian-Georgian is the aorist -e ending: b-i-Sinax-e "I kept", ele-b-i-svar-e "I whittled smth for myself", b-i-qCan-e "I turned white", b-i-Suv-e "I got wet", b-i-GurZul-e "I ate my fill/was sated". We are confident that in-depth study of Megrelian-Laz the History of the Interrelationship of Laz and Megrelian of Rioni (Abasha district villages: Gagmakodori, Guleikari, Ketilari), to the south the river Pichori separates Megrelian dialect from Gurian one.
Over the last thirteen years spreading area of Megrelian has not changed. The proofing and considering facts on Megrelian (and Megrelians) are given in Georgian hagiographic texts "Martyrdom of David and Constantine", which describes the invasion of Arabs (Marwan II -Georgians called him "Marwan the Deaf") in west Georgia and provides valuable information on the location of country of Megrelians. The text informs: "aGiZra da daibanaka qalaqsa JixanquJisasa, queKanasa megrelTasa, sanaxebsa yKondidisa, romel ars megruliTa eniTa muxa didi, da rameTu iKo banaki maTi cxeniswKliTgan vidre afxazeTamdis " -"And he encamped in the town Jikhankuji, country of Megrelians, country of Chqondidi, which in Megrelian means 'a big oak tree'. And their camps covered the area from Tskhenistsqali to Apkhazia" (Hagiographic texts, Tb., 1971, p. 258). Even today the country of Megrelians covers the area from Tskhenistsqali to Apkhazia (like during the Arab invasion in the 8 th century, in 737-741).

Conclusions
Finally, we can sum up all aforesaid, and conclude that Megrelian and Lazian demonstrate a common system at all linguistic hierarchy; they are characterized by peculiar monolithic features; they are completely of Kartvelian language type and early influence traces of other languages group aren't observable; the limit of difference between the described linguistic units indicates to their only dialectal differentiation, and therefore, based on the linguistic data, the described subsystems are qualified as two dialects -Lazian and Megrelian of one language -Zan (Colchian) one.