Caucasian Albanian and the Question of Language and Ethnicity

: Superficially, a relatively broad array of relevant data is available for statements concerning features of ethnicity with respect to the so-called ‘Caucasian Albanians’. Since antiquity, this exonymic term has been used to label both a distinct tribal group and the members of the later-on Christianized ‘Albanian kingdom’ that has existed in the northern and central parts of pre-sent-day Azerbaijan from the 2 nd to the 7 th century AD. However, little is known about the ethnic background of these Albanians. In my paper, I will pose the question to which extent the available sources allow us to make corresponding statements at all. In this context, the discovery of Bible passages written in a particular script (the so-called Caucasian Albanian palimpsests) and represen-ting a language akin to present-day Udi (a small East Caucasian minority language in Azerbaijan) plays a crucial role in determining the relevance of language in the context of the reconstruction of possible ethnic patterns underlying ancient denominations of peoples. I will argue that it is far from being self-evident to assume the existence of an ethnic unit because of a distinct language, just as a distinct language does not necessarily hint at a distinct ethnic unit. It brought about an intensification of the political and cultural influence of Iran in Armenia. Whatever the sporadic suzerainty of Rome, the country was now a part – together with Iberia (East Georgia) and (Caucasian) Albania, where other Arsacid branched reig-ned – of a pan-Arsacid family federation. Culturally, the predominance of Hellenism, as under the Artaxiads, was now followed by a predominance of “Iranianism,” and, sympto-matically, instead of Greek, as before, Parthian became the language of the educated. 33


Introduction
In book VI of his Naturalis historia (77 AD), Plinius the Elder addresses (among others) ethnographical topics of the regions of Western Asia. In Chapter 15, he talks about peoples living at the banks of the "Caspian and Hyrcanian Sea", which roughly corresponds to the region of eastern Transcaucasia. Here, he states: as people say; that part of the sea which lies in front of them, bears the name 'Albanian'. This nation, which lies along the Caucasian mountains, comes down, as we have previously stated, as far as the river Cyrus, which forms the boundary of Armenia and Iberia. Above the maritime coast of this and the Udini, the Sarmatae, the Uti, the Aorsi, (and) the Aroteres stretch along (its shores), and in their rear the Sauromatian Amazons already spoken of. 1 In this section, Plinius mentions several groups of people (Udini, Sarmatae, Uti, Aorsi, Aroteres, and Albani), one of them (the Udini) being explicitly described as a populus of the Scythians. However, Plinius does not mention any defining parameters he would have applied in order to delimit these groups from each other (or from others). In other words: What had been those social or ethnic features that marked off the concepts underlying terms like Udini, Aorsi and so on? And: Did Plinius' terms reflect patterns of ethnicity (in which sense so ever) from the internal view of the given people or just from an external view?
All we can safely state for many such terms (e.g. Aorsi, Aroteres, just to mention two of them) is that they are seemingly grounded in the construction of ethnicity on the basis of some real or fictitious features unknown to us. We cannot even be sure that the corresponding term actually represents an entity related to modern concepts of ethnicity. Starting from many of the Classical sources we might think of a very simple pattern: A group X is defined by those people who live in region Y. However, in case more than just one group is named in the context of a particular region (e.g. supra maritima eius Vdinorumque gentem Sarmatae, Vti, Aorsi, Aroteres praetenduntur quoted above) no further direct clues are given to identify the relevant features. Still, in case an ethnonym given in ancient sources (E1) is continued by a contemporary ethnonym (E2), one might tentatively argue that the people subsumed under E1 would constitute an ethnic group just as those included in E2. This might be the case, for instance with the ethnonym Μόσχοι mentioned e.g. by Herodotus . 2 It is said to refer to a part of the "[p]opulations of the coast of Pontus between the Thermodon (Sagmatas) and the Phasis in Colchis (…)", 3 cf. Herodotus: "To the Moschoi and Tibarenians and Macronians and Mossynoicoi and Mares three hundred talents were ordered (…)." 4 Admittedly, Herodotus also mentions some further details: "The Moschoi had wooden caps upon their heads, and shields and small spears, on which long points were set (…)", 5 but we can hardly refer to this information in terms of the description of an ethnic feature. 6 Nevertheless, the term Μόσχοι is usually said to be continued by the modern "Meskhetian" (Georgian meskhebi) referring to different layers of the population of the Meskheti region in Southern Georgia. 7 The problem, however, is that the meaning of the term 'Meskhetians' does not have a clear profile: It may denote Turkish speaking communities of Muslim faith just as Georgian speaking groups of Christian faith, or a mixture between both. Hence, while it may be intriguing to relate the modern term 'Meskhetian' etymologically to the term Μόσχοι, this does not necessarily mean that the Meskhetians represent remnants of the Μόσχοι of Herodotus' times.
The question of which kind of ethnicity is implied with an ethnonym of Classical times at all can be answered tentatively only, if the relevant sources add information that can be regarded as mirroring typical features of ethnicity. One example is the following passage from Herodotus, referring to the "Budinoi" (IV.108), said to have lived north of Black Sea and north of the Scythians at the river Tanais (Don). The passage is worth being quoted in details, because it nicely illustrates some of the parameters used by Herodotus to characterize an ethnic group: 8 The Budinoi are a very great and numerous race, and are all very blue-eyed and fair of skin: and in their land is built a city of wood, the name of which is Gelonos, and each 4 Μόσχοισι δὲ καὶ Τιβαρηνοῖσι καὶ Μάκρωσι καὶ Μοσσυνοίκοισι καὶ Μαρσὶ τριηκόσια τάλαντα προείρητο -Herodotus III.94. English translation Macaulay (1890): "To the Moschoi and Tibarenians and Macronians and Mossynoicoi and Mares three hundred talents were ordered (…)." 5 Μόσχοι δὲ περὶ μὲν τῇσι κεφαλῇσι κυνέας ξυλίνας εἶχον, ἀσπίδας δὲ καὶ αἰχμὰς σμικράς -Herodotus VII.78. 6 One might argue that the allusion to particular patterns of clothing etc. reflects their interpretation as a marker of ethnicity, also cf. the Suebian Knot described as a "national peculiarity" by Tacitus: "[I]nsigne gentis obliquare crinem nodoque substringere: sic Suebi a ceteris Germanis, sic Sueborum ingenui a servis separantu" ("A national peculiarity with them is to twist their hair back, and fasten it in a knot. This distinguishes the Suevi from the other Germans, as it also does their own freeborn from their slaves", cf. Furneaux (1894), Church et al. (1942)). However, it is impossible to say whether Herodotus simply describes a feature related to the appearance of the Μόσχοι that might have been present among other groups, too, or whether he viewed the "wooden cap" as a distinctive feature deliberately used by the Μόσχοι. 7 See Wimbush & Wixman (1975) for details. 8 See Thomas (2000) for the contextualization of Herodotus' Histories in contemporary intellectual discourse. side of the wall is thirty furlongs in length and lofty at the same time, all being of wood; and the houses are of wood also and the temples; for there are in it temples of Hellenic gods furnished after Hellenic fashion with sacred images and altars and cells, all of wood; and they keep festivals every other year to Dionysos and celebrate the rites of Bacchus: for the Gelonians are originally Hellenes, and they removed from the trading stations on the coast and settled among the Budinoi; and they use partly the Scythian language and partly the Hellenic. The Budinoi however do not use the same language as the Gelonians, nor is their manner of living the same (…). 9 Here, Herodotus refers to bodily appearance ("very blue-eyed and fair of skin"), to architecture ("city of wood", "each side of the wall is thirty furlongs in length and lofty at the same time, all being of wood", "the houses are of wood also and the temples"), multicultural patterns (a certain kind of religious liberalism with respect to the immigrated Greeks, cf. "for there are in it temples of Hellenic gods furnished after Hellenic fashion with sacred images and altars and cells" (…), "they keep festivals every other year to Dionysos and celebrate the rites of Bacchus"), and to language ("[The Greeks] use partly the Scythian language and partly the Hellenic. The Budinoi however do not use the same language as the Gelonians"). Finally Herodotus makes a general distinction between the Gelonian Greeks and the Budinoi by saying: "Nor is their manner of living the same (…)." Such as description -as vague as it may be -still illustrates that Herodotus tries to follow a certain canon of general features of ethnicity. This canon is best represented by his well-known delimitation of the "Greeks": (…) there is the bond of Hellenic race, by which we are of one blood and of one speech, the common temples of the gods and the common sacrifices, the manners of life which are the same for all (…). 10 Hence, Herodotus speaks of four relevant domains: (1) common genealogy, (2) common language, (3) common religious belief, (4) common habits. 11 How-9 [1] Βουδῖνοι δὲ ἔθνος ἐὸν μέγα καὶ πολλὸν γλαυκόν τε πᾶν ἰσχυρῶς ἐστι καὶ πυρρόν· πόλις δὲ ἐν αὐτοῖσι πεπόλισται ξυλίνη, οὔνομα δὲ τῇ πόλι ἐστὶ Γελωνός. τοῦ δὲ τείχεος μέγαθος κῶλον ἕκαστον τριήκοντα σταδίων ἐστί, ὑψηλὸν δὲ καὶ πᾶν ξύλινον, καὶ αἱ οἰκίαι αὐτῶν ξύλιναι καὶ τὰ ἱρά.
[2] ἔστι γὰρ δὴ αὐτόθι Ἑλληνικῶν θεῶν ἱρὰ Ἑλληνικῶς κατεσκευασμένα ἀγάλμασί τε καὶ βωμοῖσι καὶ νηοῖσι ξυλίνοισι, καὶ τῷ Διονύσῳ τριετηρίδας ἀνάγουσι καὶ βακχεύουσι. εἰσὶ γὰρ οἱ Γελωνοὶ τὸ ἀρχαῖον Ἕλληνες, ἐκ τῶν δὲ ἐμπορίων ἐξαναστάντες οἴκησαν ἐν τοῖσι Βουδίνοισι· καὶ γλώσσῃ τὰ μὲν Σκυθικῇ, τὰ δὲ Ἑλληνικῇ χρέωνται. -Herodotus IV.108. 10 (…) αὖτις δὲ τὸ Ἑλληνικὸν ἐὸν ὅμαιμόν τε καὶ ὁμόγλωσσον καὶ θεῶν ἱδρύματά τε κοινὰ καὶ θυσίαι ἤθεά τε ὁμότροπα (…) -Herodotus VIII.144. 11 Also cf. Munson (2014). ever, it remains unclear to which extent Herodotus restricted the application of these features of ethnicity to the ascription of 'Greekness' 12 or whether he regarded them as a tool for determining patterns of ethnicity in general. 13 The problem becomes more complicated out of two reasons: (a) Herodotus' characterization of patterns of -as we would call it today -the sociocultural representation of ethnicity is a mixture between an etic view reproducing information about such patterns and an emic view referring to information that may have originated from 'insiders'. (b) The mentioning of certain features by Herodotus in order to characterize a given 'ethnic unit' may be related to the narrative function of the Histories that would aim at serving the audience's expectations, namely to satisfy its demand of thrilling and exotic stories. In fact, many ethnic groups described by Herodotus and other historiographers of Classical times remained mere 'objects' without having the means and power to state features of their ethnicity in their own terms. Still, in some cases, it seems possible to retrieve relevant information from vestiges of social, economic, and cultural practices, such as cultural artifacts, settlement patterns embodied in corresponding archeological sites, or remnants of tools and clothing. 14 Nevertheless, such information may tell something about the corresponding practices of a socially and economically defined group of people. Whether or not the given patterns reflect elements of ethnicity with respect to those people who were engaged in these practices is difficult to decide. However, even if we are able to relate such findings to a particular group of people in terms of cultural representation, 15 we are faced with the problem whether the given group can be related to a documented ethnonym that would label the corresponding ethnic patterns. Furthermore, it is difficult to say whether certain cultural or economic patterns actually were part of a given ethnic identity from an emic point of view.
The global parameters set up by Herodotus in order to define "Greekness" (see above) include one point that is crucial to the topic of this paper, namely 'language'. However, while this factor seems to be important for Herodotus with respect to the question of the Pelasgians, 16 he occasionally only refers to 'language' as a parameter of 'ethnicity' elsewhere in the Histories. For instance, he says about the "Indians": "Now there are many tribes of Indians, and they do not agree with one another in language." 17 The "Androphagoi" are said to have "a language of their own" (γλῶσσαν δὲ ἰδίην, IV.106), and the "cave-dwelling Ethiopians" (τρωγλοδύται Αἰθίοπες) "use a language which resembles no other, for in it they squeak just like bats". 18 On the other hand, when coming to the Colchians and the Egyptians, Herodotus observes that "the two nations are like one another in their whole manner of living and also in their language". 19 Obviously, 'language' is seen here as an irrelevant factor with respect to the constitution of a "nation". This also holds for the characterization of some smaller units such as the "Sagartians", an Iranian group, 20 said to be "Persian in race and in language and having a dress which is midway between that of the Persians and that of the Pactyans". 21 In these short introductory remarks, I have mainly referred to Herodotus in order to illustrate some points relevant for the tentative ascription of features of "ethnicity" to social units represented by a common ethnonym. Nevertheless, much of what has been said here can be tentatively related to other ethnographers of Classical times, too. This holds especially for the feature of 'language'. In Plinius Naturalis historia, for instance, this feature is rarely referred to in order to stress the ethnicity of a given group. Rather, Plinius follows the overall dictum 'other peoples, other languages'. Thus, when talking about the city of Dioscurias in Colchis, he states, "that according to Timosthenes, 300 nations with different languages had settled". 22 Likewise, he refers to language in order to judge upon the origin of a group, as in the following passage: "[The fact] that the Celts origin from the Celtiberians in Lusitania becomes apparent γικὸν ἅμα τῇ μεταβολῇ τῇ ἐς Ἕλληνας καὶ τὴν γλῶσσαν μετέμαθε (…) "(…) if one must pronounce judging by these, the Pelasgians used to speak a Barbarian language. If therefore all the Pelasgian race was such as these, then the Attic race, being Pelasgian, at the same time when it changed and became Hellenic, unlearnt also its language." 17 ἔστι δὲ πολλὰ ἔθνεα Ἰνδῶν καὶ οὐκ ὁμόφωνα σφίσι -Herodotus III.98. 18 γλῶσσαν δὲ οὐδεμιῇ ἄλλῃ παρομοίην νενομίκασι, ἀλλὰ τετρίγασι κατά περ αἱ νυκτερίδες -Herodotus IV.183. 19 καὶ κατὰ ταὐτά, καὶ ἡ ζόη πᾶσα καὶ ἡ γλῶσσα ἐμφερής ἐστι ἀλλήλοισι -Herodotus II.109. 20 Cf. Eilers (1987). 21 ἔθνος μὲν Περσικὸν καὶ φωνῇ, σκευὴν δὲ μεταξὺ ἔχουσι πεποιημένην τῆς τε Περσικῆς καὶ τῆς Πακτυϊκῆς -Herodotus VII.85. 22 ut Timosthenes in eam CCC nationes dissimilibus linguis descendere prodiderit -Plinius VI.15. from (their) religion, language, (and) names of cities that are distinguished with the help of bynames in Baetica." 23 In sum, it seems that the dimension of 'language' generally played a minor role when attributing 'ethnic' features to a group of people. 24 Hence, it seems wise to respect Jonathan Hall's statement, according to which "[l]anguage cannot be used as an objective definition of ethnic identity". 25 Obviously, ancient ethnographers already were somehow aware of this fact. In the following section, I want to pursue the underlying question with respect to one ethnonym, namely "(Caucasian) Albanians" 26 mentioned for instance in Arrian's Anabasis, namely in his report on the battle of Gaugamela (331 B.C.): "Atropates commanded the Medes, with whom were arrayed the Cadusians, Albanians, and Sacesinians." 27 Given the fact that Arrian wrote his Anabasis in the second century AD, we cannot exclude the possibility that the mention of the "Albanians" is an anachronism. Nevertheless, if we include data from Plinius and others, 28 we can assume that in terms of an ante quem the ethnonym "(Caucasian) Albanians" had become known to the Classical world in the first century AD. However, who were these "(Caucasian) Albanians" in the sense of ethnicity?

Albanians, Udis, and Gargarians
The name "Albania" (Latin Albania, Greek Ἀλβανία, Old Armenian Ałowankʿ 29 , Parthian Ardān, Middle Persian A(r)rān, Georgian Rani, Arabic ar-Rān) alludes to a region of Eastern Transcaucasia, the core of which can roughly be associated with the Northern Azerbaijan regions left of the river Kura, 30 cf. map 1: 23 Celticos a Celtiberis ex Lusitania advenisse manifestum est sacris, lingua, oppidorum vocabulis, quae cognominibus in Baetica distinguntur -Plinius III.10. 24 Also cf. Haarmann (2014). 25 Hall (1997) 22. 26 The term "Caucasian Albanian" is generally used to distinguish the region at issue from that of Albania on the Balkans, which may have caused confusion in late Classical and early Medieval times already (cf. Ptolemaeus' mention (III.3.13.23) of the Ἀλβανοί and of a town Ἀλβανόπολις, probably located in the southern of Illyria). 27 Μήδων δὲ ἡγεῖτο Ἀτροπάτης· ξυνετάττοντο δὲ Μήδοις Καδούσιοί τε καὶ Ἀλβανοὶ καὶ Σακεσῖναι -Arrian Anabasis 3.8.4, also cf. 3.11.4; cf. Robson (1967). 28 Also cf. e.g. Tacitus Histories 1.6, Tacitus Annals 6.35, Plutarch, Pompeius 35; Cassius Dio, Historia Romana 37.3-4. 29 All quotes from Armenian sources are taken (together with their English translation) from Gippert et al. (2009). See this edition for fuller reference. 30 See Trever (1959), Bais (2001), Gippert et al. (2009). The map refers to "(Caucasian) Albania" in terms of a political unit ("kingdom"). In fact, representatives of the region had become relevant political players and rulers since the end of the 2nd century BC, probably in the context of the wars between the Arsacid Mithridates II and the Armenian king Artavazd I. 31 Strabo describes the emergence of the Albanian 'state' as follows: "Their kings, also, are excellent. At the present time, indeed, one king rules all the tribes, but formerly the several tribes were ruled separately by kings of their own according to their several languages." 32 Else, nothing precise is known about this process that ended up in a more or less stable political unit being around 70 BC. Nevertheless, after Pompeius had defeated the Albanian king Oroezes (or: Orodes) in 65 BC, Albania became a vassal or protectorate of the Roman empire, still under obvious cultural impact from Parthia. This vassalage lasted roughly 300 years, interrupted by periods of closer ties to the Parthian empire. But the consolidation as a more or less independent political unit ended under the rule of the Arsacids. What has been described by Toumanoff for Armenia grosso modo holds for Caucasian Albania, too: 31 Marquart (1901); Trever (1959) 149. 32 διαφέρουσι δὲ καὶ οἱ βασιλεῖς: νυνὶ μὲν οὖν εἷς ἁπάντων ἄρχει, πρότερον δὲ καὶ καθ᾽ ἑκάστην γλῶτταν ἰδίᾳ ἐβασιλεύοντο ἕκαστοι. -Strabo, Geography 11.4.6. Translation according to Jones (1924).
It brought about an intensification of the political and cultural influence of Iran in Armenia. Whatever the sporadic suzerainty of Rome, the country was now a part -together with Iberia (East Georgia) and (Caucasian) Albania, where other Arsacid branched reigned -of a pan-Arsacid family federation. Culturally, the predominance of Hellenism, as under the Artaxiads, was now followed by a predominance of "Iranianism," and, symptomatically, instead of Greek, as before, Parthian became the language of the educated. 33 With respect to Caucasian Albania, this process became even more pronounced after the region was subdued by the Sassanid regime (252-253 AD). Nevertheless, although Albania became a Sassanid satrapy, it still kept its local Arsacid dynasty for nearly 150 years. One of the rulers, namely King Uṙnayr of Albania was baptized by Gregory the Illuminator at about 314 AD, and he subsequently declared Christianity as the official religion in Albania (lasting until the 8 th century). From that time on, the history of the regions of Albania was dominated by two factors: (a) by the political relations to Armenia and the Sassanid empire, (b) by the local Church history. It is beyond the scope of this paper to recapitulate the milestones of this history the description of which mainly dwells upon Old Armenian sources. 34 The major sources are Pat-mowtʿiwn Ałowanicʿ ašxarhi, 35 Koryun's Varkʿ Maštocʿi "The Life of Mashtots", 36 Anania Širakacʿi's Ašxar hacʿoycʿ, 37 and Movsēs Xorenacʿi's Pat-mowtʿiwn Hayocʿ. 38 Hence, we are somehow informed about the ecclesiastic and political structure of the statehood of Albania from the 4th century AD onwards. Still, the question remains whether we can relate this statehood to a particular ethnic unit termed "Caucasian Albanians". This question has become of special importance since the discovery of the so-called "Caucasian Albanian Palimpsests". 39 These in parts heavily damaged palimpsests had been found in the Catherine Monastery on Mt. Sinai in the year 1975 and identified as for their lower layer as "Caucasian Albanian" by Zaza Aleksidze in the years 1996-2001. The two relevant manuscripts (Sin.Geo. N 13 and N 55) could be identified as a part of an early Christian lectionary and as a fragment (nearly a half) of the Gospel of John. 40 These texts (in sum roughly 13.000 readable tokens) were written in the so-called Caucasian Albanian alphabet, akin to the world of Georgian and Armenian scripts, but different in shape, phonetic value, and ordering of the individual signs. The fact that this script was termed "Caucasian Albanian" is mainly grounded in the fact that it in parts corresponds to an alphabet documented in the Yerevan manuscript Mat. 7117. In this manuscript from the 15th century, the presentation of the alphabet is preceded by the words "The script of the Albanians is: […]". 41 The identification is further corroborated by a report (1307)  And he [Mesrop] inquired and examined the barbaric diction of the Albanian language, and then through his usual god-given keenness of mind invented an alphabet, which he, through the grace of Christ, successfully organized and put in order. (…) And when this order was actually fulfilled and bore results, Bishop Eremia soon began the translation of the divine books, whereby at once in one instant, the barbaric, slothful, and brutal men became well acquainted with the prophets and the apostles, becoming heirs to the gospels, and in no way ignorant of the divine traditions. 43 In the somewhat later account by Movsēs Xorenacʿi and in the Patmowtʿiwn Ałowanicʿ ašxarhi, however, the name of the people has changed from "Albanians" to "Gargarians": And he called a certain Beniamin, a gifted translator whom Vasak the lord of Siwnikʿ had delivered as a child into the hands of Anania, his bishop, (and) who created the letters of that guttural, harsh, ugly and barbarous language of the Gargarians. 44 41 Աղուանից գիրն է։, cf. Abuladze (1938), Gippert et al. (2009).
A third ethnonym / toponym attributed to the world of Caucasian Albania is documented by the set Udini, Uti, Οὔτιοι, Otene resp. Ὠτηνή, and Ուտի (Owti) resp. Ուտի առանձնակ (Owti aṙanjnak "Uti proper"). The terms denote either a 'people' (Udini, Uti, Οὔτιοι) or a region/province (Otene resp. Ὠτηνή, Ուտի (Owti). The fact that these terms have to be under consideration in the context of the 'Albanians' is grounded among others in the following observations: (a) As has been said in the first section of this paper, Plinius the Elder mentions the Udini as well as the Utii as tribes located near the "Albanian Sea" or in "Albania", cf. again Plinius: At the entrance, on the right hand side, dwell the Udini, a Scythian tribe, at the very angle of the mouth. Then along the coast there are the Albani, the descendants of Jason, as people say; that part of the sea which lies in front of them, bears the name 'Albanian.' This nation, which lies along the Caucasian mountains, comes down, as we have previously stated, as far as the river Cyrus, which forms the boundary of Armenia and Iberia. Above the maritime coast of this (nation) and the (nation of the) Udini, the Sarmatae, the Uti, the Aorsi (and) the Aroteres stretch along (its shores), and in their rear the Sauromatian Amazons (which we have) already spoken of. 49 (b) In his Ašxarhacʿoycʿ (see above note 33), Anania Širakacʿi mentions a province Owti said to be controlled by the Albanians: The thirteenth (province), Uti, in the west of the Arax (river) between Arcʿax and the river Kura, has seven cantons, which are held by the Albanians: Aranṙovt, Tṙi, Ṙotptak, Aławe, Tučʿ katak, Gardmun, Šakašēn, (and) Uti proper, where the city of Partaw (is). And it has olive oil and cucumber, and Jasmine, and there is also the katak of the birds. 50 (c) Movsēs Kałankatowacʿi also relates the tribe of the Utis (ագզ Ուտեացւոց) to the country of Alowankʿ: On the establishment of his rule over the northerners, he (sc. Vałaršak, king of Armenia) summoned to him the wild, foreign tribes in the northern plain and round the foot of the Caucasus and in the valleys and ravines south thereof down to the entrance to the plain, and commanded them to cease their plundering and murdering and to pay tribute to the king. He appointed over them governors and prefects of whom the chief, by order of Vałaršak, was a certain Aṙan of the Sisakan family, descended from Japheth, who received the plains and mountains of Albania from the river Araxes to the fortress of Hnarakert. And they called the country Aluankʿ on account of the sweetness of his ways, for they called him ału ['sweet'] on account of his agreeable disposition. And from among his descendants, they say, famous and valiant men, many governors were appointed by Vałaršak the Parthian; and from his son, they say, descended the inhabitants of the principalities of Uti, Gardman, (Covdkʿ, and Gargarkʿ). 51 (d) On March 20, 1724 a group of 'Udi' people from Sheki/Nukha (NW-Azerbaijan) sent a petition written in Armenian to Tsar Peter I asking for support and protection against the "unlawful ones and unbelievers". Here, they identify themselves as follows: "(…) We are Albanians (ałowankʿ) and, by parentage, Udis (utikʿ). By the preaching of the apostle Ełišē our forefathers became believers of God (…)." 52 Taking these wordings literally, the term owtikʿ "Udis" would denote the genealogical identity of the ałowankʿ ('Albanians'). Referring to the second sentence, ałowankʿ would then simply mean 'local Christians' in the tradition of the apostle Ełišē. 53 In sum, there are three terms that are associated with the present problem, namely the question of the ethnic identification of the 'Albanians': Alban-, gargar-, and uti-/udi-. Whereas alban-and uti-are explicitly used as toponyms, too, gargar-seems to be restricted to its semantics as an ethnonym (in its broadest sense). However, the question remains whether these terms used as ethnonyms refer to the same group or people at all.
53 According to local traditions and in reference to cf. Kałankatowac ʿ i (I.6), Ełišē (Elisaeus of Albania) was the first patriarch of the Church of Caucasian Albania. He was said to have been a disciple of St. Thaddeus (or, of Thaddeus of Edessa), again one of the Twelve Apostles of Jesus.
Ἀλβανία, Old Armenian Ałowankʿ, Parthian Ardān, Middle Persian A(r)rān, Georgian Rani, Arabic ar-Rân) the most likely source is an Old Iranian form *aldwān or *ardwān 54 that has not undergone the 'Parthian' shift dw > b). As far as data go, this underlying form (aldwān or *ardwān) cannot be linked convincingly to other etyma. 55 When looking at Greek Ἀλβανία, we have to consider the suffix -ία that derives (among others) toponyms from ethnomys or the like. This would in fact suggest that Ἀλβαν-had been an ethnonym (cf. Ἀλβανοί). The Armenian term աղուանք Ałowankʿ is ambiguous in this respect. It is used to denote both the collective of 'Albanian people' and the region. Taking -kʿ as the Armenian plural morpheme, we may conclude that ałowan-itself meant an ethnonym. Given the clear Iranian background of this term, its derivation from Armenian աղու (ałow) 'soft, sweet, agreeable, affable' as suggested by Movsēs Kałankatowacʿi (I, 4) seems to be a folk etymology. Nevertheless, one might think of an analogous process in Parthian or Middle Persian. Parthian Ardān and Middle Persian A(r)rān would then contain the plural marker -ān, added to a stem *ardw-or *aldw-. It is tempting to relate this element to e.g. ardu-in the Old Iranian proper name Ardumaniš, perhaps meaning 'upright minded'. 56 The underlying form *arda-may be related to "the PIE *h 2 erd-'high, lofty' that otherwise appears in Latin arduus 'steep, towering, lofty', Old Irish ard 'high, great', Hittite harduppi 'a high' (…)". 57 In the sense, ard(w)-ān would then have denoted something like 'heights' or 'high regions' (referring to regions from the left bank of the river Kura to the Great Caucasus).
The term Udi-/Uti-is likewise obscure. The two variants may reflect a phonetic unit perceived as -d-or -t-in the different sources. Most likely, we have 54 Cf. Gippert et al. (2009) for a fuller discussion; Parthian Ardān should then be regarded as a loan from another Middle Iranian language (substituting earlier *a r / l ban-?). 55 Terms like alban, designating a group of Dom people in Syria (Meyer (2004) 74) or Alpan, the name of a small and municipality in the Quba Rayon of Azerbaijan have to regarded as false friends. In Azerbaijan, it is nevertheless very current to relate the toponym Alpan to Albania first suggested by the Azerbaijani writer Abbas-Kuli-aga Bakixanov (1794-1847) (cf. Bakixanov (1991) ch. 1,60: "In the province of Quba, there is a village Alpan, the name of which is probably derived from Alban(i)a", tr. W.S.), even claiming a Turkic origin of the term, also cf. Zardabli (2014) 55: "Although some researchers consider that, the name of »Albans« is not understandable, this word and calling of »Alban«, »Alpan« is referred to Turkish background. Basing on the ancient Turkish word of »alp//alb« (brace, courageous, daring, plucky, manly), we can say that, the name of Albania means »the country of braves, the state of men«" (typ. errors etc. in the original). Zardabli's statement actually reflects the official doctrine of the Azerbaijani State, as the book was published "by order (…) of the Ministry of Education of the Azerbaijan Republic". -Perhaps, the Middle Persian A(r)rān variant has survived in Sorani Kurdish aran 'fertile land'. 56 Cf. Tavernier (2004) 12. 57 Adams (1994) 12. to deal with a tense voiceless dental stop (-t:-), which allows reconstructing a form *ut:i-. Nevertheless, one has to admit that Plinius' mentions of the Udini and the Uti as distinct groups (relating the first to the Scythians) speaks against an identification of both. It is quite remarkable that the terms Otene (Ὠτηνή) 58 and Uti (Arm. Owti) denote regions rather than an ethnic group. It is not evident, however, that Otene (< *utene?) has been derived from uti. The nasal segment in Plinius' udini might likewise reflect the Latin derivational element -in-forming (among others) ethnonyms. If this is correct, we would have to separate the name of the province (Otene) from the term udi-/uti-. This would also mean that the two toponyms Otene and outi(kʿ) refer to two different regions. If this is true, the safest claim concerning the pair udi-/uti-is that it represents an old toponym (lowlands west of the river Kura extending to the eastern slopes of the Artsakh/Karabagh mountains in the west and to the river Araxes in the south, 59 cf. map 2: Map 2: Possible location of Albania, Uti, and Gargar in the regions of Azerbaijan.
The question remains why today the off-springs of the overall 'Albanian' tradition, namely the Udis in the villages of Nij and Vartashen/Oguz in northern Azerbaijan name themselves udixo ∼ udiux (plural). Today, the Udis, Christians by belief, mainly settle in the village of Nij (Azerbaijani Nic, Udi niˤž), located in Northwestern Azerbaijan and inhabited by some 6.000 people. In 2009, 58 It should additionally be noted that in Ptolemaeus' map (liber geographiae 5.12; tertia Asii tabula), the core of the Utik ʿ region is labeled Totene (but not Otene). 59 See Gippert et al. (2009) for this localization. some 65 % of the inhabitants of Nij declared to be ethnic Udis, the rest being chiefly Azerbaijanis. Nij is divided into sixteen 'family-based' quarters (šaq'q'a or mähällä), two of which are mainly inhabited by Azerbaijanis (Yalgaşlı, Abdallı). Until 1989, a more or less compact group of ethnic Udis was present in the village of Vartashen (now Oğuz), too, located some 20 km northwest of Nij. Before 1989, Vartashen had been inhabited by some 5.000 people (roughly 40 % Armenians, 15 % Jewish Tats, and 30 % Udis). Together with the local Armenians, most of the Udis from Vartashen were forced to leave the village in 1990 due to the Armenian-Azerbaijani conflict and thus moved to various places of the former USSR. The village of Zinobiani (1938 in Eastern Georgia had been founded by emigrants from Vartashen in 1922 in the context of the Armenian-Azerbaijani conflict 1918-1920. As has been said above, the first reference to the existence of 'Udis' north of the river Kura dates back to 1724 (see the quote from a letter of the people in Sheki given above). Prior to this date, the regions north of the river Kura were generally known as Albania. Still, there is some evidence that the Udis once maintained stronger ties with the northern regions of historical Uti (Tawush). Still in the 18th century, the orientation of Udis towards Tawush (and Qarabagh) seems to have played a crucial role in every-day life in Nij and Vartashen, as illustrated by the following passage from Schiefner's report on the Udis (compiled at about 1855): "Dabei wuchsen die Bedrückungen von Seiten der Chane, besonders stark verfolgte Mahmed Hassan Chan (1783-1804) die christlichen Uden, welche zum Theil aus Wartaschen und Nidsch nach Qarabagh flüchteten (…)." 60 The fact that today the Udis name themselves udi-is perhaps related to the adaption of the ethnonymic tradition in the former Uti region. Unfortunately, there is no information about what the endonym of the Udis north of the Kura (that is in historical Albania) had been in medieval times. The only indirect evidence is given by some Arabic sources: According to Arabic geographers such as al-Muqaddasī (∼ 985) people spoke (by that time) "Albanian" (ar-rāniyya) in the city of Bardaʿa (Partav), also labeled the 'Mother of Rān' (ummu 'r-rāni). Al-Istakhrī (ante 957) notes: The language of [the people] of Azerbaijan, of Armenia and of Arrān is Persian and Arabic, except for the people of Dabil and its surroundings who speak Armenian, and [the language of] the land of Bardaʿa is Albanian. 61 60 Schiefner (1863) 5. This orientation seems to be vivid still today. After the exile of most Udis from Vartashen/Oghuz (1989/90) in the context of the Armenian-Azerbaijani conflict, some refugee families took shelter in the Armenian Tawush region (Bagratashen, Ptghavan, Debetavan, and Haghtanak), see Schulze & Schulze (2016) 513-535. 61 De Goeie (1927) 191-192, my translation. The Christian layer in Bardaʿa is reflected by the fact that according to al-Muqaddasī "Sunday was the market day in Bardaʿa, called 'Kuriku' (Kürikü)". 62 Still, it is a matter of dispute whether this language was 'Albanian' at all. Instead, one might assume that we have to deal with a local Northwest Iranian variety.
Nevertheless, the use of Armenian ałowan-to denote the Udi language seems to have been current until the 19th century, as illustrated by the title of the manscript Načal'nye osnovanija grammatiki na Agvanskom jazyke, pisannoj armjanskimi bukvami. (ms., St. Peterburg Oriental Institute, no. C-7 (Armenica) "composed by an Udi from Dashbulaq, in the province of Sheki". 63 In 1893, the Vardapet Makar Barxowtarean published a description of the Ałowanicʿ erkir ('Albanian region') together with an Armenian-Vartashen Udi word list that comprises 209 words and one phrase. 64 Prior to the official recognition of the Udis as being descendants of the Albanians, systematically propagated among the Udis, local traditions that would relate the Udis to the Albanians in terms of ancestry were not widely known. Nevertheless, Schiefner (1893, 57-58) documents a passage "on the Udis" probably told by a certain Stephan Bezhanov (a local teacher in Vartashen): The people from Vartashen and my father say that once there was an Udi kingdom and that the city of the throne was Vardan (i.e Bardaʿa) (…). How it came, how it went, how the Udi kingdom was destroyed -the Udi do not know. In Chamchean the Armenian's history 65 it is written how the Udi kings led war against the kings of the Armenians, (how they) were defeated, (how they) paid taxes, how they rose again, (how they) defeated the Armenians (and how they) became tranquil (i.e. independent) for some times. 66 62 I.e. Greek Κυριακή; see Karaulov (1909) (1827). History of Armenia. Translated from the original Armenian by Johannes Avdall, in two volumes. Calcutta: Townsend). Note that Č ʿ amč ʿ ean never uses the term 'Albanian' or 'Albanians'. Instead, he refers to the province of Uti and its people. 66 Translation W.S. (Vartašlun amdarġon next'un, bez bawanal nexe te udiġoy padšaxluġ bakene, t'axtey sähäral Vardanney (…). Hetära hario, hetära tacio, hetära pučbakio udi padšaxluġ, šet'ġox udiġon tet'un ava. Arminġoy Čamčiin ist'oriin boš came, hetärt'un udiġoy padšaxġon arminġoy padšaxġoxon davabio, čaxecio, xarj tadio, hetärt'un p'urum ini haizerio arminġoy čaxnio, ičġoy vaxt'in dinjaliś bakio).
Summing up what has been said so far, it seems reasonable to assume that the three terms at issue are ultimately derived from older (perhaps descriptive) toponyms, which somehow matches the construction of many ethnonyms in ancient times. The terms thus denoted three different regions in the same area before the term 'Albanians' acquired a broader reading because of the growing regional supremacy of Albanian rulers. Nevertheless, in case the denominations 'Albanians', 'Utis/Udis', and 'Gargarians' firstly refer to the people living in these regions, it is not self-evident that these people would have shared corresponding patterns of ethnicity or that they would even have shared a common ethnic identity. In order to pursue this question, it is relevant to single out relevant sociocultural features, traditions, and norms shared by the individual groups.
3 Ethnicity, language, and the peoples of Albania

Ethnicity
As has been alluded to in the introductory section, there seems to exist a certain automatism in correlating ethnonyms of Classical times with features of 'ethnicity'. In certain cases, there is sufficient evidence to corroborate corresponding assumptions. 67 With respect to the topic of the present paper, a major problem is given by the fact that we have mostly to deal with external views that report rumors and saying rather than are grounded in more or less objective observations. The following table lists the key information of the most prominent sources (leaving aside geographical information as well as Strabo's account (XI.4.1-8, see Tab. 1). The table illustrates that we can hardly associate the information of the corresponding author to parameters of ethnicity as listed e.g. by Haarmann. 68 Here, the author names the following six macro-categories:

a) Descent b) Constituents of human ecology (7 features) c) Sociocultural markers (5 features) d) Communication systems (5 features) e) Interaction and social behavior (2 features) f) Phenomenological markers (7 features)
It is not the intent of the present paper to discuss and evaluate these categories together with their corresponding features in details. Nevertheless, it should be stressed that we have to deal -as Haarmann puts it -"with a theoretical construct. Each of the features may be of significance for shaping a group's identity, but not all the features (…) will have equal significance in each instance". 69 In fact, an ethnic unit is basically a social unit that construes and maintains a collective identity based on a bundle of distinctive features present in the traditional and conventionalized sociocultural knowledge patterns of the community. Each group defines the relevance of the corresponding features differently. The way people do this is again part of the cultural knowledge of the community. Nevertheless, we have to bear in mind that cultural knowledge usually is tacit and part of the individual's "conjunctive knowledge" in terms of Mannheim (1980Mannheim ( [1922), that is as "implicit, experiential, non-reflective, praxeological knowledge grounded in everyday practices". 70 It would be of importance to have available a larger corpus of artifacts, products and so on that would document these practices. The presence of long-standing patterns observable with these artifacts would then hint at a possible relevance for the group at issue, which again would be relatable to the internal perception of the group's 'ethnicity'. As for Caucasian Albania, such data are hardly available. 71 Religious buildings and their architecture do not necessarily express 'ethnocultural' peculiarities. Rather, they may document a given religious profile the symbols of which were introduced and controlled by institutional organs. Only in case we have clear evidence that a group relates its own religious practices to these buildings, we might assume that their form part of the group's ethnic identity. This aspect leads to the assumption that the 'ethnicity' of a group becomes more apparent in case its administration has become institutionalized: In this case, people might refer to the given institutions as the embodiment of their own ethnicity (in which way colored so ever). Still, it would be problematic to turn the evidence around by saying that if there is information about a person that ruled over a regionally defined population, the ethnic affiliation of the ruler must have been that of the whole population under control. The same does not hold only for 'rulers' as such, but also for the relevant ruling classes.
69 Haarmann (2014) 22. 70 Schulze (2016) 190. 71 except for some few every-day objects the provenience of which, however, is not also ascertained, see Aslanov & Vaidov & Ione (1959), Trever (1959), Rzaev (1976 From an external view, only few data are available that relate to sociocultural patterns of the peoples at issue. The main source is Strabo . This passage relates to reports on the 'Albanians' stemming from the first century BC. Accordingly, the economic patterns of the 'Albanians' were dominated by sheep-breeding, horticulture, and agriculture (using wooden ploughs), and viniculture. They had no money and no measures (Strabo mentions that they do not use numbers over '100'). 72 Referring to demographic issues, Strabo says that the Albanians were able to supply an infantry of 60.000 men together with 22.000 horsemen. Taking these data serious, we would have to infer that the whole population of 'Albanians' by that time was around 400.000 people, which seems rather exaggerated. According to Strabo, men were tall and beautiful, having a straight attitude, but were not very martial. In sum, they led a "Cycoplean mode of life". 73 In these pre-Christian times, the 'Albanians' were said to worship three main gods, namely the Sun, Jupiter (Zeus), and the Moon. According to Strabo, the Moon was their most important god. 74 The priests played a central role in the 'Albanian' society and came next to the king in authority. The 'Albanians' knew human sacrifices and forecasting related to the way a person killed by a sacred lance fell down. With respect to their social structure, the 'Albanians' seemed to be organized in different units headed by chieftains and marked for different languages. In the first century BC, these different units came under the control of a single 'king' (114.6). If we take Strabo's observation literally, according to which "[t]hey speak six and twenty languages from the want of mutual intercourse and communication with one another", 75 we may assume that the early statehood of Albania was populated by people from twenty-six groups defined in terms of their linguistic affiliation. 76 Starting from Strabo's information, we would then have to conclude that the term 'Albanian' is not related to a single language. Rather, the term would characterize a group of people that shared the above-mentioned traditions and that dwelt in the regions of 'Albania', later on controlled by a single chieftain or 'king'.

Language
The aspect mentioned in the last paragraph of the preceding section is of special importance, when it comes to the role of language in the ascription of 'ethnic features' to the 'Albanians' (resp. Utis and Gargarians). Here, two major points have to be distinguished: (a) The role of a language in the context of communicative traditions: Starting from Strabo's description of the linguistic landscape in Albania, 'language' seems to have played a certain role in the constitution of the individual social groups in Albania (prior to Christianization). This would argue in favor of the assumption that the people of Early Albania defined their social affiliation (among others) by reference to 'their' language as opposed to the languages of the surrounding groups. In this respect it is irrelevant whether we have to deal with dialects or individual languages from a system-linguistic point of view. It makes more sense to refer to the famous dictum reported by Max Weinreich: "a shprakh iz a dialekt mit an armey un flot" ("a language is a dialect with an army and a navy"). 77 Accordingly, a 'language' can be described as some kind of 'armed dialect', controlled and administered by corresponding official institutions. Strabo's observation according to which there had been a correlation 'group ⇆ language ⇆ king' in earlier times of Albania suggests that the individual languages were indeed languages rather than dialects from a sociolinguistic point of view. Still, there is no information about which languages were included in the linguistic landscape of Early Albania. At least one of these languages must have been the language referred to by the above-mentioned Armenian sources when talking about the introduction of a script to the 'Albanian people'. The authors characterize the corresponding language as the "guttural, harsh, ugly (and) barbarous language of the Gargarians" 78 or talk about "the barbaric diction of the Albanian language". 79 It is generally acknowledged that these famous passages refer to languages of the East Caucasian (more precisely, Lezgian) language family (mainly because of the mention of "guttural" sounds). Still, it remains unclear whether Mesrop Maštocʿ worked on just one language. The Patmutʿiwn Ałowanicʿ reports that He (…) spread the teaching of the gospel to the land of the Utiacʿikʿ, the Albanians, the Lpʿinkʿ, the Kaspkʿ, up to the Čʿołay Pass, and to other foreign tribes whom Alexander of Macedon had captured and settled around the great Mount Caucasus, namely, the Gar-garkʿ and the Kamicik Hepʿtʿalkʿ (Hephthalites); he reconverted them to the Christian 77 Weinreich (1945) 13. 78 Movsēs Xorenacʿi, Patmowtʿiwn Hayocʿ III,54; Patmowtʿiwn Ałowanicʿ ašxarhi II,3. 79 Koryun,Varkʿ Maštocʿi XVI (288). faith and taught them the form of worship which they had learned long ago and had now forgotten. A perfect preacher and apostle to the barbarous mountain tribes, he taught them to write in their own language. 80 The question is whether "their own language" is meant in the sense of "language of each people", or whether it indicates that the Utiacʿikʿ, Albanians, Lpʿinkʿ, Kaspkʿ, Gargarkʿ, and Kamicik Hepʿtʿalkʿ all spoke the same language. Taking Strabo's account literally, we would have to assume that these groups had been marked for different languages. Nevertheless, there is no evidence that peoples like the Lpʿinkʿ (Latin Lupeni), Hepʿtʿalkʿ, or Kaspkʿ (Caspians) ever used a script (also see fn. 76). The matter is further complicated because the source given above mentions the 'Albanians' as a unit opposed to those of the Utiacʿikʿ, the Lpʿinkʿ, or the Kaspkʿ. Obviously, 'Albanian' did not serve as a cover term in the sense of Strabo, but rather as a separate ethnonym. It hence remains an open question, whether there had been a group of 'Albanians' that differed from the other peoples in the region by language.
(b) The function of language products (texts, etc.): As has been said above, the Armenian sources report about the introduction of a script to the 'Albanians' or 'Gargarians' (created by Mesrop Maštocʿ around 422 AD). The purpose was to use this script for translating the Bible. Koryun tells us: Bishop Eremia soon began the translation of the divine books, whereby at once in one instant, the barbaric, slothful, and brutal men became well acquainted with the prophets and the apostles, becoming heirs to the gospels, and in no way ignorant of the divine traditions. 81 This passage is of special importance because it suggests that by that time, Albania was inhabited mainly by people who did not know Armenian or Georgian. The dating, however, is problematic: Remnants of the Caucasian Albanian Bible texts (fragments of the Gospel of John, fragments of a lectionary) as preserved in the Caucasian Albanian palimpsests from Mt. Sinai suggest that the text had been produced on the basis of (among others) an Armenian translation. 82 Given the fact that the first Armenian translation had been done between 411 AD and 434 AD, and that Bishop Eremia seems to have been a contemporary of Mesrop Maštocʿ (362-440), we have to assume that the process of translating the divine texts into Albanian (or Gargarian), perhaps just initiated by Eremia, stretched over a longer period in the middle of the 5th century. 83 The whole enterprise seems to have been brought about by the Albanian king Arsuał(ēn) (or: Esualen). It is reasonable to assume that Mesrop, Eremia, and his interpreter Benjamin referred to the language of the court when aiming at the translation of the Bible. In this sense, we can safely start from the hypothesis that in the 5th century, the language showing up in the Caucasian Albanian palimpsests had been the language of the court of Albania, that is the language of the ruling elite. However, it would be premature to infer from this hypothesis that the whole population of Albania spoke this language. Referring to Strabo's account quoted above, we may assume that the political unit 'Albania' included a number of diverse groups marked by different languages. Although one should bear in mind that Strabo's account talks about a situation some 500 years before the introduction of the 'Albanian' script, there is no reason to claim that with these 500 years, Albania would have undergone full homogenization regarding its linguistic landscape. On the other hand, we have to consider the fact that the Bible translation had obviously been used for propagating Christian belief. This becomes visible especially from the lectionary included in the palimpsests, cf. Figure 1.
The letters of the lectionary texts are somewhat larger than those of the Gospel manuscript. This goes together with the general function of lectionaries, namely to be read out during church services. This again means that the audience should have at least understood the language. However, this does not necessarily mean that 'Albanian' was part of the ethnic patterns of the 'Albanians'. First, we have to acknowledge the fact that the eastern regions of Caucasia had since long been marked for functional multilingualism, a pattern which is preserved in some regions till our days. Second, the audience may have had a passive knowledge of 'Albanian' activated just in the context of church services. Third, it remains unclear who attended these services. It may well have been the case that mainly (or even only) the 'Albanian' elite took part in such events.
82 Gippert et al. (2009). 83 The Caucasian Albanian texts as documented in the palimpsests seem to have been produced in the 7th century. Most likely, we have to deal with a copy of an older original.  (Gippert et al. 2009, Vol. II, VII,31).
Another question related to the 'ethnic relevance' of Caucasian Albanian concerns the degree to which literacy might have been given in this kingdom. 84 The fact that only 400 years later, the script seems to have fallen into oblivion suggests that literacy was not fully anchored in the society of Albania. Some of the reasons are probably related to the fact that after 510 (…) the confessional and ecclesiastical evolution of the country steadily furthered the Armenian tongue and script at the expense of the Albanian. (…) These confessional debates and religious struggles are likely to have once for all abolished what had remained from the Biblical collections and the liturgical books compiled in the Albanian tongue during the Vth-VIth centuries. Since the criterion of orthodoxy had become the official doctrine of the Armenian Church, any attachment to local peculiarities or, even worse, any use of books in a non-Armenian tongue might have been regarded as suspect and been severely repressed. 85 84 For the question of literacy in Classical times, cf. among others Harris (1989), Thomas (1992), Bowman and Woolf (1994), Lomas, Whitehouse, and Wilkins (2007), Johnson and Parker (2009). 85 Gippert et al. (2009) I, xviii-ix. Again, this probably held especially for the use of the Caucasian Albanian script by local intellectuals and clergies. It is difficult to decide to which extent the script had also been used in non-religious or profane contexts. An indication may be the presence of six short inscriptions on potsherds and candleholders, cf. e.g. Figure 2.
However, documents of this kind are too few to allow drawing a more comprehensive picture on the presence and relevance of the Caucasian Albanian script in a private context. Things are somewhat different with respect to the famous inscription on the Mingečaur pedestal, 86 cf. Figure 3.
Most likely we have to deal with a pedestal of a cross. The text indicates that the inscription had been prepared in the year 558 AD (reference is probably made to Xosrow I. Anōšērwān (531-579)). A tentative transliteration and translation is: 87 and …åy instead of … ås (4a). Gippert also convincingly argued that the reading biyayn<e> 'he made' given in Gippert et al. (2009) should be given up in favor of biyayn. This short inscription illustrates that the Caucasian Albanian script was also part of the linguistic landscape of Albania. It hence had a presentational function, too, and might even be seen as an expression of public communication in all its modesty. Summing up this point, we may conclude that the Caucasian Albanian script had been part of the religious and perhaps political culture of the Caucasian Albanian elite. Up to now, there is no evidence that the script had been used for more than one language. Admittedly, not all inscriptions can be safely read on the basis of our knowledge of the language of the palimpsests. However, this may be due to the fact that the religious texts (to the extent they are readable at all) contain a lexicon that may not or only in parts be present in the relevant inscriptions.
on the basis of the scant data available by those times. This concerns not only the Mingečaur pedestal (see above), but also a list of month names which "is contained in the encyclopedic works of the authors Anania Širakacʿi and Hovhannes Imastaser 89 and further appears in Georgian manuscripts containing the lexicon of Sulxan-Saba Orbeliani". 90 This list, analyzed in details by Gippert (1988), entails some names that can best be interpreted in terms of Udi (e.g. *ex-n'-ay '(month of the) harvest', cf. Udi ex 'harvest'). This list is also of importance because it illustrates that a highly relevant segment of everydaylife organization, namely the naming of months in the context of the calendar system, was carried out in Caucasian Albanian (albeit we have to admit that at least some of these names are Iranian). Still, the bulk of evidence for the Albanian-Udi hypothesis stems from the texts of the palimpsests. 91 Accordingly, we have to assume that the language of those people who used the language of the palimpsests was an ancestor of present-day Udi. Most likely however, we are not dealing with an immediate ancestor of one of the presently two varieties of Udi (Nij and Vartashen), but rather with an 'aunt' standing in slightly closer relation to actual Nij Udi than to Vartashen Udi. 92  Given the present location of the Udis (who once covered a much broader region north of the river Kura), we might assume that the differentiation between 'Caucasian Albanian', 'Proto-Nij', and 'Proto-Vartashen' had taken place in the first centuries AD. Still, we can hardly determine the region related to this process. A possible scenario would be that the two varieties of Caucasian Albanian and Proto-Nij had resulted from a migration to the south, even beyond the river Kura. Evidently, at least the speakers of Caucasian Albanian came into closer contact with Iranian groups. 93 After members of those people speaking Caucasian Albanian gained control over the region, they seemed to face the problem of becoming fully iranized with respect to language. Most likely, the process of Christianization contributed to the fact that such an Iranization did not happen at least until the end of the 5th century. The use of the 'Albanian' language in religious contexts as well as its codification in terms of the Bible translation probably served as a barrier against language shift, at least as long as Parthian resp. Middle Iranian were concerned. In this sense, we can assume that 'Albanian' functioned as a marker of religious and maybe even political identity, strongly associated with the local autocephalous Church. This does not necessarily mean, however, that 'Albanian' also was a factor of ethnicity, if ever we can relate the ruling class of Albania to an ethnic unit at all. 94 After the Arsacid royal house of Albania (a branch of the ruling dynasty of Parthia) had become extinct in 510 (death of Vačʿagan III Barepašt), it was replaced by princes of the Persian or Parthian Mihranid family (assuming the Persian title of Arrānšāh 'King of Albania') who controlled the region until Varaz-Tiridates II (821-822). We cannot tell which language the nobles of Albania used by that time, but it seems rather likely that they mainly spoke Middle Persian, at least in an official context. Most likely, 'Albanian' became reduced to a vernacular, also because the more the Albanian Church came under control of the Armenian Orthodox Church, the more Albanian lost ground in favor of Armenian. 'Albanian' thus gradually lost its function as an element of religious identity. It is reasonable to assume that this process did not necessarily affect the ethnic patterns of those who used to speak 'Albanian' before. The cultural and social traditions of the local people probably were not largely affected by the ongoing language shift. Summing up we can say that the 'Albanian' language represented mainly a particular communicative system perhaps associated to a group of people in the regions of Albania members of which gained power over other groups. For some times (roughly 400-600), the language became part of the cultural capi-93 although we have to accept that Iranian had been a general factor of impact on all varieties of Udi, see Schulze (2015b) for details. 94 It should be noted that so far, none of the proper names given for the kings, nobles, and clergies of Albania shows any affinity with what we know about Caucasian Albanian. Mostly, we have to deal with Armenian and/or Iranian names. tal of the ruling class, 95 if not of the Albanian speaking population as such. During this relatively small span of time, Albanian might have functioned as a marker of ethnicity, although it is more likely that it mainly served as a factor of social and religious identification. In this sense, it is difficult to claim that the givenness of 'Albanian' suggests an 'Albanian ethnicity'. When looking at the linguistic data available from the palimpsests, only very few features become visible that would hint at cultural or even ethnic peculiarities of the speakers of the language. Obviously, this is mainly due to the nature of the texts. These texts importantly show to which degree the translators referred to Armenian, Georgian, Greek, and Syriac sources, 96 but we can hardly claim that the adoption of foreign phrasings and words (be it directly, be it in terms of loan translations) would necessarily speak in favor of a corresponding knowledge of the audience or readership. Only in case we are dealing with terms that reflect common concepts we may assume that they are grounded in more general patterns of language contact, cf. CA avel 'many, much' (cf. the Armenian root *awel-in aweli, aṙ-awel 'more'), the abbreviation žd corresponding to Armenian žołovowrd 'crowd, people, synagogue', CA garazman 'grave, sepulchre' (Arm. gerezman 'id.'), CA mowˁak' 'worker' (cf. Arm. mšak, Georg. mušak'-i 'id.'), just to name a few. The following diagram classifies the loans from Armenian, Georgian, Iranian, and Syriac according to relevant domains: As one might expect, the religious domain constitutes the main body of loans in the language of the palimpsests. The other domains do not indicate a clear profile that could be related to particular forms of cultural impact on the society of Albania. In fact, such loans mainly tell about patterns of communicative interaction of the 'Albanians' with Armenian, Georgian, or Iranian speaking people. They can hardly be used as indicators of particular cultural patterns associable to some kind of 'Albanian' ethnicity. 52,8 % of the lexical stems documented in the palimpsests (n = 570) can be safely related to Udi (40,8 %) or to other East Caucasian languages (12,8 %; the remaining 33,8 % are of unclear origin). If ever cultural peculiarities were mirrored in the language of the Caucasian Albanian palimpsests, we would have to search for them within the set of lexical units that Albanian shares with Udi, but that does not have matches in other East Caucasian languages. Such innovations (archaisms are less likely) might in parts reflect cultural innovations related to times when Proto-Udi had emerged. Likewise, the 73 lexical stems that escape a secure etymology up to now perhaps might be another clue to this point. Admittedly, etymological research on the lexicon of Udi and Caucasian Albanian still is in its beginnings. 97 However, even if we were able to classify the relevant data in terms of parameters of Cultural Linguistics, we would not be allowed telling with certainty that the underlying concepts had played the role of key concepts expressing features of ethnicity by that time. In order to trace possible correlations, one would have to refer to non-linguistic evidence reflecting the cultural relevance of the corresponding concepts. Unfortunately, such data are not available yet.

Summary
Superficially, a relatively broad array of relevant data seems to be available for statements concerning features of ethnicity with respect to the 'Caucasian Albanians'. In fact, the assumption that the 'Albanians' represented an ethnic unit in Antiquity seems to have been rather current since long. However, the 'ethnic' identity of the Albanians has rarely been discussed in details yet. Referring to the available data one might think that a closer inspection of the assumed ethnic properties of the 'Albanians' would be rather promising: We have at hands a broader set of geographic information (Herodotus, Plinius, Strabo, Ptolemaeus etc.). Strabo gives a quite extensive ethnographic description of the Albanians. The Armenian sources supply us with detailed information on the political and ecclesiastic history of the Albanian Kingdom, and, last but not least, the discovery and decipherment of the Caucasian Albanian palimpsests allows describing in much details the language of the 'Albanians'. Nonetheless, when trying to use all this information in order to draw a coherent picture, it soon becomes evident that a patchwork of arguments and observations emerges instead of the expected coherent picture. This unsatisfactory result is due to many factors some of which I have tried to discuss in more details above. The main problem seems to be that up to now most of what can be said about the Albanians either stems from early sources reflecting external views (or even rumors) on the 'Albanians' or is doubtful to a certain extent. What is even more important: We cannot be sure that the different sources (covering a period of some 500 to 700 years) talk about the same topic at all. For instance, the identification of Strabo's Ἀλβανοί with the Ałowankʿ of the Armenian sources mainly dwells upon the ethnonym itself. However, we cannot ascertain that the two terms refer to a group of people sharing a common history (not to speak a common 'ethnic' identity). Admittedly, the localization of the 'Albanians' in some regions of present-day Northern and Central Azerbaijan that is common to most relevant sources may argue in favor of this assumption. Nevertheless, we cannot exclude that the denomination 'Albanian' had been applied to different groups in that region at different times.
In this respect, we might question whether the term 'Albanian' represents an ethnonym in its true sense at all. First of all, the name seems to refer to people living in the regions of 'Albania' (just as it would hold for the terms 'Udi' and 'Gargar'). This seems to have been the case in Strabo's account. In later times, after a kind of fiefdom had been established in the region, the term might have simply denoted those people who were put under the authority of the corresponding rulers. The first attribution of a collective identity to this people seems to be related to Christianization. However, albeit religion can surely serve as an important marker of societal identity, it can hardly be related to ethnicity as such. In a sense, one might think of relating the fact that Albanian Christianity developed into an autocephalous structure to the notion of 'ethnic religion' or 'indigenous religion', 98 but the period of religious independence was much too short for allowing the stabilization of the Christian belief in Albanian in this respect. In addition, Armenian sources tell us that Christianization in the Eastern (Trans-)Caucasus was not confined to the 'Albanians'.
Hence, we might likewise assume that Christian identity in the regions of Albania was not confined to the group of 'Albanians' in early medieval times.
As has been said above, language is often seen as a typical marker of ethnicity. In the case of the Albanians, we have to acknowledge, however, that we cannot even tell for sure that the available written documents (the palimpsests and the inscriptions) are 'Albanian' at all. Nowhere in the readable parts of the palimpsests we find an indication that would say something like *alowanowġoy mowzen campêne 'it is written in the language of the Albanians'. The only indication that the text indeed is 'Albanian' comes from indirect sources, e.g. from the name of the alphabet given in the alphabet list and from the list of month names referred to above. 99 Still, we cannot exclude the possibility that the script called 'Albanian' in the alphabet list had been used for languages other than Albanian, too. In this respect, we have to remember that Movsēs Xorenacʿi (Patmowtʿiwn Hayocʿ III.54) talks about the Gargarians for whom Mesrop Maštocʿ had developed a script. Up to now, we cannot say whether the Gargarians were part of an 'Albanian confederation' sharing with the Albanians the same language. All we can say for sure is that people speaking the language of the palimpsests must have been present in the kingdom of Albania.
It goes without saying that the language of the Caucasian Albanian palimpsests conventionally called 'Albanian' played a considerable role at the times of early Christianity in present-day Azerbaijan. Nevertheless, it is premature to assume that it marked a particular group with respect to specific features of ethnicity. The popular view, according to which a language necessarily reflects an ethnic unit is likely to fail at least with respect to the question of the Caucasian Albanians.