Ana Kharanauli 

Questions about the Origin of the Translation of the Georgian Bible

Georgian Christian culture – literary culture, literary language, conceptual system, literary form, etc. originate from the translation of the Bible. Hence, knowledge of from which language, where, in what circle and when was the Bible translated into Georgian will by itself allow us to shed light on many general and concrete questions in the sphere of religion, culture, history of language and philology.

Study of the Georgian Bible from the philological viewpoint obviously begins with the origin of the translation. Otherwise, it is absolutely impossible to determine the text of the Bible, research into the history of the text, discussion of the character of the translation(s) and exegetic or any other type of commentary.

Identification of the original of the Georgian translation of the Bible is interesting for the study of issues connected with this original itself: to ascertain to what extent did the original of Scripture condition the legitimacy of its translation, which textual form was dominant in the Church at the time when the Georgian translation was made, what was its textual peculiarity, what semantic did it have – vocabulary and phraseology, to what extent were its new and alien elements comprehensible to a nation whose different religious views and whose language differed. Without ascertaining the specific original it will be also impossible to use the Georgian translation for research into the phenomenon of the medieval translation or for studies connected with the emancipation of the language.

Then, from which text was the Bible translated into Georgian? How concretely can we speak of the original of the Georgian translation of Scripture? Before passing on to these questions, let us make clear the concepts which we shall subsequently discuss.

What is implied under the “Georgian Bible”?

The point is that the translation of every individual book has its history; in the available MSS the translation has been corrected from different points of view, with varying quality and accordingly to different originals. Hence, the “Georgian Bible” is not a single whole collection which implies translation(s) of each separate book of the Old and New Testaments: a whole number of these books has been repeatedly translated into Georgian and their origin is linked to various sources and different literary circles.

Besides, the “Georgian Bible” implies the original translations and not the extant form that is attested in the translation that is distanced at least by five centuries.

How is the original of the translation to be established?

The traditional view on the origin of the Georgian translation which stems from the 19th and gains ground early in the 20th century, was mainly given shape by superficial views existing regarding the Church of the Christian East and social-political history. Subsequently, the provenance of the Georgian Bible turned into part of demonstration of Nicholas Marr’s theory, while identification of textological data and their interpretation fully fitted theoretical views.

Actually, however, the original of a translation must be established on the basis of intratextual and intertextual research. In particular:

1. The translation should be collated with all possible sources, i.e. MSS that have come down to us. This methodological demand was largely neglected in the case of Georgian translation and the view on its provenance took shape on the basis of comparison with the critical text of various editions.

2. Establishment of the origin of the translation and separation of the layers of the translation and recension is practically a synchronous process, implying the study of the translation technique (which implies both the specificity of a particular translation and the general trend established over the history of Georgian translation) and consideration of the developmental history of the language. It is wrong to ascribe to an ancient translation what was not characteristic of it at a given stage of the development of the language or did not fit with the character of the original translation.

3. Texts are compared to one another according to different parameters and that are not of the same weight for the ascertainment of the provenance of the translation.

The main parameter for the establishment of the provenance of the Georgian translation is correspondence in the sphere of syntax: this primarily refers to rendering infinitive constructions and sequence of words in a sentence. Change of the sentence structure results in radical changes of the text, and revisers resort to it in the case of extremely formalistic approach to translation –later, in the Hellenophile period. Thus, the syntactic construction reflects the earliest layer of a translation.

Wrong comprehension of the original’s MS or errors stemming from wrong spelling is a good argument for the identification of the source; however, bearing in mind the typology of errors, we should treat this argument too with caution.

The quantitative data can be used as a criterion of research on the provenance of the source only with consideration of the translation technique: to ascertain whether the extension or shortening of the text, which corresponds to the foreign source, belongs to the translation or recension, we should bear in mind the tendency of translating or revising, and for this the lexical or grammatical equivalents of the added parts should be considered in relation to the remaining part of the text.

The vocabulary is an extremely multi-plane criterion in researching the provenance of a translation. In determining the relation between the translation and the source attention is usually claimed by borrowed vocabulary, which is frequently the cause of wrong conclusions. In the first place we take into consideration with what we are dealing – with old loans or borrowing as direct part of the translation process. Study of the rendering of proper names and drawing conclusions call for their own methodological approaches. Observation of the segmentation of lexical units yields a more reliable picture: segmentation of a word under the word level (i.e. breaking a word into parts and formal or semantic reflection of these parts in the translation) and creation of neologisms by lexical or semantic calquing is indicative of direct relation to the text. However, determining whether this relation is linked with the old translation or the stage of reviewing should again be done by bearing in mind the principle of the technique and reviewing. Study of the stability of lexical equivalents is obligatory element in using the vocabulary as a criterion for establishing the parent text.

Thus, a Georgian translation must be compared to its possible parent texts according to all mentioned parameters. Incidentally, each of them has much more weight when they represent the variant readings of some - Armenian or Greek - versions or separate textual groups. E.g. when the Georgian shares at the same time both the variant of the Armenian translation and the variant of the hexaplaric recension of the Septuagint, we obviously cannot define provenance of this variant in the Georgian translation. When the Georgian follows the variant that occurs only in one version, this points to a direct relation. However, here too we have an “if”: we may be dealing with a typological coincidence, e.g. a certain understanding of the text, a certain resolution of a problem (such is, e.g. in rendering an infinite construction by a finite one, or semantic translation of an idiom, or certain rendering of a concept non-existent in the language and so on). Ultimately the question is decided by a thoroughly analyzed and differentiated statistical picture (but not statistics of unanalyzed material!).

Thus, the establishment of the parent text of the translation is connected with complex research. Now let us see how perfectly it has been carried out to obtain the conclusions that are circulating in Georgian and especially in foreign textbooks.

What is implied under the parent text of a Georgian translation?

 On the one hand, the language of the parent text. Here two options are discussed: Greek and Armenian. Frequently, when speaking about the parent text, it is merely noted from which language the Georgian translation was made. In this case one or two lexical Armenisms or Greecisms are usually “sufficient”. On the other hand, a concrete textual tradition of Armenian and Greek Bible is implied. In the case of the Old Testament the Greek tradition is rather diverse, beginning with the text, and it is rewritten form and ending with the large recensions of the Septuagint (hexaplaric, lucianic, catenae) and textual groups of individual MSS. Unfortunately, Armenian textual tradition is not so well studied. Within the Pentateuch it is believed that the Zohrab Bible gives a fairly uniform picture of the MS tradition, in the case of Isaiah the MSS are distanced from one another.

How we shall present briefly the picture that has taken shape to date regarding the provenance of the Georgian translation stemming from Greek and Armenian. We shall begin with the Armenian.

 What is implied under “Armenian”?

 When the thesis on the Georgian translation stemming from the Armenian was taking shape it was believed that there existed an ancient Armenian translation, the so-called Armena 1, surviving only sporadically in quotations. It was also believed that this Armenian translation was done from an ancient Syriac text, surviving only in a later form. The Georgian text was needed for the reconstruction of these lost originals. When there was a Syriac boom in Europe interest in Georgian was due also to this – they looked for a frozen trace of a dinosaur (R. Blake, J. Molitor)

Another part of scholars (J. Assfalg, M. Shanidze) considered the actually existing Armenian as the “Armenian” text, Zohrab’s edition (text and its apparatus) is considered to be its valid representative.

On what is the view on the Georgian translation stemming from the Armenian based?

 The idea stems from the extratextual view on Christianity entering Georgia from Armenia, and accordingly the Georgian literary culture, beginning with the creation of the alphabet, took shape via the Armenian. This is the foremost methodological error – discourse on the text only on the basis of its historical context– even understood correctly.

As to the textual data, traditionally the greatest importance is attached to lexical Armenisms. At the same time the fact is not taken into account that we may be dealing with old loan words, already established in the language, which is indicated also by the fact that this word as usual does not occur in the relevant passage of the Armenian text. In addition, words borrowed in Georgian and Armenian from Pahlavi are often considered to be Armenisms. Typological coincidences in selecting lexical equivalents are also considered as lexical Armenisms (e.g. descriptive translation of “herd of goats”).

Analysis of formal correspondences is also conducted in a wrong way: account is not taken of the potentialities and linguistic preferences, e.g. that both Georgian and Armenian prefer a dependent clause as an equivalent of infinitive construction, etc.

The case is similar in quantitative correspondence: in a sentence without a predicate restoration of the predicate is more natural for both languages, or a translator into both languages resorts to one and the same translation technique: he imitates the parent text, adding a pronoun to a definite type of nouns.

The similarity of the order may also be determined by a tendency existing in the language, i.e. the nominal and verbal parts of a compound predicate should stand side by side, etc. Grammatical coincidences may also be the result of typological coincidence, e.g. general meaning is natural for Georgian and from this viewpoint Georgian-Armenian correspondences in number are irrelevant in determining the provenance of the translation.

As the witnesses of the sources are not fully taken into account, in the scholarly literature, arguments in favor of affinity with Armenian are based also on readings that are attested to by other groups of Greek MSS. Account is taken only of similarity with Georgian and differences are ignored; the relationship of the Georgian to Armenian variant readings (Sonderlesrten - which have no parallels in other sources) are not treated separately.

But even when the differences from the Armenian variant readings are obvious, and by all the above parameters the Georgian translation is appreciably inclined to the Septuagint and its recensions, it is then assumed that either the Armenian text formed the basis for the Georgian that has not come down to us, or was strongly edited according to the Greek. Even though this also appears not plausible, because such editing according to the Greek - purposeful correction of the Scripture against the Greek Vorlage is not assumed prior to Mount Athos (J. Assfalg).

Why is it believed that the text of the Georgian translation was corrected according to the Greek rather than the Armenian?

 Because it is considered doubtful that correction could have taken place according to the Armenian after the split between the Churches. But if we take into account the interests of Chalcedonian Armenians to have a reliable Scripture, the correction of the Georgian according to the Armenian must have taken place precisely after the split. This is supported also by the fact that only single MSS have been corrected according to the Armenian, e.g. the Mtskheta Psalterion, or parts of the Adishi Gospel.

In short, this is the logic of the theory of the Georgian translation of the Bible from the Armenian; if the Georgian from any point of view follows the Armenian, this is a sign that it is translated from the Armenian, and if it does not follow it, i.e. it is corrected according to the Greek, or is the only source of the oldest, lost Armenian translation and its corresponding also lost Syriac texts.

Now let us revert to the question of the interrelationship of Georgian and Armenian texts. What conclusion can one arrive at on the basis of the textual study of the books of the Old Testament?

On the basis of the Georgian translations that are more or less well studied from this point of view, we can speak of two types of interrelationship of the Georgian and Armenian translations:

 One part of the books of the Old Testament – the books of Kings, the Psalms have been corrected according to the Armenian. The correction deals mainly with the vocabulary and the quantitative changes.

As to the translation stemming directly from the Armenian, to date only two texts have been brought to light: the version of Ezra 1 and 2, preserved by the Oshki Bible. This dependence is confirmed by all textual data and is especially seen well as a result of comparison with the Greek text:

 1. The Georgian and Armenian texts follow the same Greek original, which sharply differs from Lucian’s tradition of the Septuagint. 2. The Georgian follows the Armenian and is opposed to Greek in syntactical constructions by conveying the function of Greek cases and prepositions, the word order; 3. Lexical units are conveyed in Georgian according to the Armenian: a Greek compound noun is translated by one segment and vice versa, a Greek simple word is rendered by a compound word; 4. The phraseology follows word for word an Armenian phrase translated freely from the Greek, idioms and phrases characteristic of the Septuagint are translated in the same way as the Armenian; 5. The Georgian follows the Armenian from the quantitative viewpoint as well, adding what and where the Armenian does. It also follows the omissions of the Armenian.

Thus, the text of Ezra 1 can be used as a specimen for characterizing a translation stemming from the Armenian in order to show clearly the boundary that lies between 1. what is translated from the Armenian, 2. reviewed according to the Armenian, and 3. the Georgian text that is quite independent from the Armenian.

 But the translation of Ezra 1 made from the Armenian is not the only and the oldest translation of this book. Now let us pass on to the short cut that leads to the origin of the Georgian Bible.

 What is implied under “Greek”?

 The oldest Georgian translation and Khanmeti fragments

The Khanmeti fragments (the palimpsest fragments of 5^th-7^th centuries) indeed justified the expectation created by their discovery early in the 20th century. In the first place, they enabled us to shed light on the diversity of the history of the translation of the Bible. It proved that Ezra 1 and Jeremiah’s Prophesy, whose oldest translation was known to us by the Oshki MS, is given in the Khanmeti fragments in a different translation, while Isaiah’s text of the same Oshki MS actually repeats the text of the Khanmeti fragment. At this point we are not interested in the reason of this diversity. We are now talking only of the provenance of the extant texts and are interested in ascertaining to what extent these textual forms constitute the oldest translation. From this standpoint the Khanmeti fragments are a key to the problem: they are documentary proof form of the Georgian text that was in use in 5th-7th cc.

However, before we discuss the provenance of the translation, we shall first shed light on the textual tradition the newly converted country came to face in the entire Christian Church.

Early in the 4th c. Hieronymus speaks of three mutual texts (trifaria varietas) that were current in different regions: Origen’s text in Palestine, Lucian in Antioch and Constantinople, and the text of Hesychios in Egypt. From which textual type could the Old Testament have been translated into Georgian. Theoretically none can be ruled out – neither the Hexaplaric text or the three textual forms about which Hieronymus speaks, for in the early Christian age the trace of Georgian Christians in these three regions (Palestine, Antioch-Constantinople, Egypt) is well-defined. But what has the textual study of Khanmeti fragments shown?

 The origin of the translations found in the Khanmeti fragments.

 As I have said, the Khanmeti Ezra 1 is an absolutely different translation from the Oshki Ezra 1. According to all the parameters by which the provenance of the translation is established, it is a text translated from the Ezra 1 MS of the lucianic recension. Furthermore, of the two MSS that reflect text of this recension, it is closer to one, 108th MS and occasionally presents the text form of the lucianic recension more authentically. Close relation to lucianic MSS is consistently demonstrated over two different sections: 1. In the list of proper names (Ch.9) it preserves both the forms of these names (the inaccurate readings, found in the fragment, can be explained by the confusion of Greek majuscule letters, which is also proof of the fact that the translation is made from a Greek MS) as well as its order; 2. In the narrative part of the text (Ch. 1-2) it reflects the quantitative composition of Lucianic MSS, lexical alternation, word order. In short, Khanmeti text of 1. Ezra has no trace of Armenian.

The relation of Khanmeti Jeremiah to the Jeremiah of Oshki and Jerusalem MSS and to its Greek and Armenian equivalents does not yield such a clear picture, as does a comparison of the two Georgian translations of Ezra 1 with each other and to Greek and Armenian texts. The reason of this is that both translations of Jeremiah stem from the Vorlagen (in the given text section it is hard to distinguish between the hexaplaric and lucianic texts) which were close to each other, besides, both translations use the same translation technique. For the same reason the difference of both Georgian translations of Jeremiah from Armenian is not so obvious: the Armenian too is a translation of the hexaplaric recension of the Septuagint. Therefore, the outcome is to observe what picture the Georgian translations give when the Armenian readings deviate from Greek counterparts. The comparison of the Georgian readings with the Armenian ones inclines me to assume that Vorlage of the Georgian translation was the Greek manuscript of hexaplaric/lucianic recension.

 In the case of Khanmeti Isaiah the difference from the Armenian is greater. The Georgian translation follows the lucianic recension – in my view, the oldest layer of this recension, while the Armenian is often opposed to these lucianic variant readings. At the same time the Georgian does not share the Armenian variants, neither when the latter, along with hexaplaric sources, is opposed to the lucianic (with one exception), nor its variant readings – addition, reduction, transposition, lexical changes.

 Thus, the Khanmeti fragments of the Old Testament prove unequivocally that the oldest layer of the extant Georgian translation is absolutely free from a trace of the Armenian and reflects concrete Greek textual tradition fairly well.

The origin of the so-called Oshki type translations

Next to the Khanmeti fragments, the oldest MS is the well-known Oshki Codex (978). Along with the collection of the texts of the Jerusalem Prophets (n. 7/11, 11th c.) and the Jerusalem MSS (n.113, 13th c.) it has preserved a text more or less close to the Georgian archetype, which I tentatively call Oshki type text, though this term unites neither translation of the same provenance or of the same text history.

If we continue the theme of the text of the Prophets: the Oshki – Jerusalem MSS translations stem also from the Greek. The vocabulary of the book of Amos: preverbal forms, compound nouns, rendering proper names, semantic differences, rendering specific phraseology by various techniques, different textual traditions and finally, translations’ errors stemming from the Greek – all this bespeaks of the Georgian provenance from the Greek, not attesting to any trace of Armenian. The rendering of nominal constructions, word order opposed to Armenian variant readings (Sonderlesarten) and the quantitative composition strengthen this conclusion further.

Among historical texts the books of the Kings are extremely complex in textual composition, being apparently the result of the contamination of Georgian translations of texts of Hexaplaric provenance. In some MSS the probable Armenian influences still await study. 

The study of the origin of Solomon’s books further enriches our knowledge of the provenance of the Old Testament: e.g., Ecclesiastes reveals a connection with the text of the early Egyptian Text Group (the Papyrus 998 and Codex Sinaiticus).

Two versions of Esther are translated into Georgian; they too are directly based on the Greek parent text. The Oshki translation is done from the version of Esther that is known under the name of alpha text, and whose Armenian equivalent is unknown.

The provenance of the so-called apocryphal texts of the Old Testament is unequivocal: e.g. the book of Tobit is translated from the so-called intermediate version of the Septuagint (LXX^III), differing from the Armenian, Baruch from the Lucianic recenzion, while the Armenian evidences affinity rather with the Hexaplaric recension.

The Georgian translation of the Psalms should be singled out, whose oldest layer is preserved in Sinaitic MSS and is fairly stable from the textual viewpoint. It is related to the Greek lucianic manuscripts and so-called Psalterium Gallicanum of Latin Psalms. This Georgian translation, that prior to Giorgi the Athonite’s recension was the Vulgate for the Georgian Church, has been corrected according to the Armenian in the so-called Mtskheta MS: the text corrected according to the Armenian is written on the erased words.
This brief summary of research conducted over the last decade into the issue of origin of the Georgian Bible translation is obviously not enough to dispel the skepsis or ignorance that exists in the West regarding the Georgian translation. The purpose of the paper was to bring to the fore the methodological questions related to research into the Georgian translations and to draw attention to the new literature appearing recently in the scholarly press.