Focused Series »

Indo-European Origins
Northern California
The Caucasus
Imaginary Geography
Home » Cultural Geography, Historical Geography, Linguistic Geography, Nationalism, Russia, Ukraine, and Caucasus, The Caucasus

Historical Clues and Modern Controversies in the Northeastern Caucasus: Udi and Ancient Albania

Submitted by on January 23, 2012 – 5:19 pm 32 Comments |  
Map of Hurrian Kingdoms and Neighbors, Circa 2300 BCEThe Caucasus is rightly called a “mountain of languages.” Linguistic diversity reaches its extreme in the Russian republic of Dagestan and adjacent districts in northern Azerbaijan. The nearly three million inhabitants of Dagestan speak more than thirty languages, most of them limited to the republic. Such languages may seem inconsequential to outsiders, mere relict tongues of minor peoples. Yet a few of them are of historical significance, and the broader linguistic geography of the region provides evidence of important historical patterns stretching back for thousands of years. Historical linguistic relationships here are also implicated in the current-day political struggles of this troubled region.

Although the languages of Dagestan include members of the widespread Turkic and Indo-European families, most belong to the Northeastern Caucasus family. Many linguists include the Nakh languages of neighboring Chechnya and Ingushetia in the group; others essentially limit it to languages spoken in Dagestan and the mountains of northern Azerbaijan. Despite its restricted distribution, the NE Caucasian family is deeply differentiated, including six clearly separate subfamilies in addition to Nakh. Three of these groupings (Lezgic, Dargin, and Avar-Andic) include one or two “major” languages, spoken by hundreds of thousands of people, along with an assortment of local tongues used by only a few thousand. According to the Wikipedia, four Dagestani languages (Avar, Dargwa, Lezgin, and Tabasaran) are “literary,” employed to some extent for written communication.

As a remote area with many small ethnolinguistic groups, the northeastern Caucasus is distinctive but hardly unique. Other areas of forbidding topography with similar levels of linguistic diversity include the highlands of New Guinea—a vastly larger area—and the Nuba Mountains of Sudan. Most such areas are assumed to be historical backwaters, but that is not the case in regard to the northeastern Caucasus. If the “Alarodian” hypothesis is correct, two of the most important  peoples of the ancient Near East spoke languages, now long extinct, that were closely linked to the Northeastern Caucasian family.

The ancient languages in question are Hurrian and Urartian. “Hurrian” may not be a household word, but various Hurrian states were rivals of the Babylonians, the Hittites, and other Bronze-Age “super-powers.” The Hittite Empire itself probably included large numbers of Hurrian-speakers, although its official language was Indo-European. The main body of the Hurrians, living in what is now northern Iraq, northeastern Syria, and southeastern Turkey, also seems to have been over-run by Indo-Europeans, chariot-riders who established the powerful Kingdom of Mittani circa 1500 BCE. The Mitanni rulers had Indo-European names, but they soon adopted the Hurrian speech of their subjects, as revealed by the remarkable Amarna Correspondences preserved in Egypt.

Map of Ancient UrartuThe Mitanni Kingdom of the Hurrians disappeared in the conflagration that marked the end of the Bronze Age, circa 1200 BCE, a time of massive population movement, de-urbanization, and the retreat of literacy. By the tenth century BCE, however, a powerful new kingdom using a closely related language emerged in the area around Lake Van in what is now eastern Turkey. This Iron-Age kingdom of Urartu was noted for its mineral wealth and for its bitter rivalry with the Assyrian Empire. Urartu persisted until it was conquered by the Empire of the Medes, the immediate predecessor of the Persian Empire, circa 590 BCE. At roughly the same time, the land of Urartu seems to have been linguistically transformed by the spread of proto-Armenians from the west, a people perhaps linked with the ancient Phrygians who spoke a language in an outlying branch of the Indo-European family. In the twentieth century, Armenian nationalists began to glorify ancient Urartu as the deep font of Armenian culture. In doing so, they sought to highlight the antiquity of their claims to territory in what is now eastern Turkey. Without endorsing such political claims, it is only fair to acknowledge a close historical connection between Urartu and Armenia.

Map of Albania in the Caucasus and Neighboring Kingdoms, Circa 300 CEThe linkage between NE Caucasian languages and ancient kingdoms is strongest in Caucasian Albania, a state that covered much of what is now Azerbaijan from the fourth century BCE to the eighth century CE. Like Armenia and the Georgian kingdom of Iberia, Albania was politically caught between, and deeply influenced by, the Persian world to its east and the Greco-Roman world to its west. We know from ancient Greek writers that the Albanians eventually acquired their own script, but knowledge of this writing system was lost until 1937. At that time, a Georgian scholar discovered a reproduction of the Albanian alphabet in a medieval Armenian manuscript. Subsequently, a few stone inscriptions were found that used the same script, but the language itself basically remained a mystery until the early 2000s.

Map of the Kingdoms of the Caucasus Circa 300The story of the recovery of Albanian writing begins in 1975, when a fire damaged a number of manuscripts in a neglected basement cell in the famous Eastern Orthodox monastery of Saint Catherine’s in Egypt’s Sinai Peninsula. The heating of the manuscripts helped reveal the fact that some were palimpsests, parchment manuscripts that had been scraped over and then re-inscribed. Fifteen years later, unknown letters were noticed under a Georgian text in one of the documents. In 1996, the Georgian scholar Zaza Alexidze determined that the underlying passages were in Albanian. After several years of concerted effort, he recovered and translated the entire hidden layer of the palimpsest. What he found was an Albanian Christian lectionary, a church calendar with specific scriptural readings keyed to specific dates. Some scholars believe that this long-forgotten and thoroughly erased text, which dates to the late forth or early fifth century, is the oldest Christian lectionary in existence.

Map of NE Caucasian Languages, Including UdiAlexidze’s translation was facilitated by the existence of a living tongue strikingly similar to the language used in the lectionary. The literary language of the ancient Albanians, it turns out, lived on among the Udi, a group of eight thousand persons inhabiting two villages in Azerbaijan. As the years passed, the Udi language diverged from old Albanian, but not by much. The surviving Udi people also retained the faith of their ancestors. Although they live in a largely Muslim area, the modern Udi belong to their own Udi-Albanian Christian church.

Christianity originally spread to Albania from Armenia. The Albanian church eventually separated from the Armenian, affiliating instead with the Orthodox Christianity of the Greek world. After the Muslim conquest of Albania in the 600s, such an affiliation became politically fraught, as the Greek-speaking Byzantine Empire was the main principal rival of the Muslim Caliphate. As a result, the Albanian Christian population was again placed under the ecclesiastical authority of the Armenian Apostolic Church. Over time, it seems, much or perhaps most of the Albanian population assimilated into the Armenian community. Those who resisted Armenian religious control seem to have evolved into the modern Udi. Yet the Udi population continued to decline, as many members adopted Islam and were absorbed by the Azeri community. Today, the Udi language is regarded as gravely endangered.

As might be expected, the Albanian heritage of the eastern Caucasus has generated a contemporary political controversy among Armenian and Azerbaijani partisans, focusing on the disputed territory of Nagorno-Karabakh. Eastern Armenians, according to some Azerbaijani stalwarts, are not so much genuine Armenians as transformed Albanians—like much of the Azeri population. Armenian scholars charge Azerbaijani historians with greatly exaggerating the extent of Albanian assimilation, and with trying to “de-Armenianize” much of the historically constituted Armenian region.

To the neutral bystander, the issue might seem moot; ethnic groups and nations often expand by assimilation, and the mixing of peoples is more the norm than the exception over the long term. Primordialist nationalism, however, retains a strong hold on the imagination, especially when faced with intractable military conflicts. As the “frozen war” between Armenia and Azerbaijani is now going into its third decade, it is not surprising that the ancient Albanians would be recruited into the conflict.

Previous Post
Next Post

Subscribe For Updates

It would be a pleasure to have you back on GeoCurrents in the future. You can sign up for email updates or follow our RSS Feed, Facebook, or Twitter for notifications of each new post:

Commenting Guidelines: GeoCurrents is a forum for the respectful exchange of ideas, and loaded political commentary can detract from that. We ask that you as a reader keep this in mind when sharing your thoughts in the comments below.

  • I did not know that Udi is modern “Albanian” (and there lies the Serbian nationalist hypothesis of modern Albanians being arrivals from the Caucasus based only on the name).

    More so, I find interesting that Udi has not changed that much since 2000 or 1600 years ago. I am of the opinion that non-expanding isolated languages evolve much more slowly than expanding cosmopolitan ones (because of lack of creolization and such). This would seem to support this idea.

    In the end you have not really addressed the issue of whether Hurro-Urartean is related to NE Caucasian or not. I was hoping for some clarifying analysis on this matter (on a comment by Asya in the previous thread).

    • Sorry if I’ve misled you: I did not know the precise content of Martin’s post,only the general topic.

      As for your comment on the rate of change of different types of languages, it turns out that smaller languages without literacy change much faster than big standardized languages, which makes the claim about slow-changing Udi suspicious. What I think happened here is that *written forms* of ancient Albanian and Udi look similar, sort of like Biblical and Modern Hebrew do, in writing.

      • That seems to contradict what we know about Icelandic (in the context of Scandinavian languages) or Lithuanian (in the context of IE languages). I can see how widespread literacy (and TV and what not) may delay change by imposing standards but that’s such a modern anomaly that it’s mostly irrelevant for historical linguistics.

        • Icelandic and Lithuanian are but two examples, both somewhat exceptional. Worldwide, small languages of tribal groups (note that neither Icelandic nor Lithuanian fit that description) change with dizzying speed. Without writing (again, something that both Icelandic & Lithuanian had) Udi didn’t have a chance to stay as still as the scholar who deciphered ancient Albanian would have us believe.

          As for the standardizing role of TV and internet and so on, it shouldn’t be forgotten that religious books often had that role since antiquity. Good examples include King James Version of the Bible, which contributed significantly to the standardization of Early Modern English in the 16th century or the role of the Hebrew Bible as a standard for later versions of Hebrew.

          • There was no writing in Lithuania until a few centuries ago (first book is from 1547), while Lithuanian conservatism is relative to Indoeuropean as a whole, which is maybe 6000 years old. I imagine that most Lithuanians were illiterate until some time after the Bolshevik Revolution, what is like “yesterday”, or at least well in the 19th century (and this happened in most countries on Earth: widespread literacy is a most recent phenomenon).

            I think you’re pushing some ideas of yours without really thinking about them.

            Hebrew and Jewish literacy(which, for what I have read, was rather common, unlike what happened with the rest of ethnicities) are so exceptional that is not worth considering except in its own specific frame. Also modern Hebrew has been reinvented from the ashes. If you’re imagining the reality of languages based on Hebrew, which is little more than a dead religious language artificially resuscitated for the purposes of Zionism, then it’s easy to explain how you can reach to such strange conclusions.

          • Maju, I base my claims on research that has been done on Australian Aboriginal languages, which are said to have change unrecognizeably in about a 100 years. Just for comparison, we can still read Shakespeare in the 500-year old English.

            As for the Hebrew revival, you might find the series of posts on this topic interesting:

            As for the issue of literacy, you missed my point: one doesn’t have to be able to read (a printed book, no less!) to be exposed to a standardized form of language. Nowadays, one can simply turn on the TV, and in the older days, one would go to church (or some other form of religious worship)…

          • In 100 years?! I just don’t believe that: there MUST be an error. If it’s not an error then there is a most anomalous phenomenon (like massive creolization by influence of English or something like that). A language can’t change as fast as making communication with your great-grandfather impossible just because of such change!

            As for Hebrew: “Hebrew gradually became extinct as a spoken language around 200 CE”. That’s my point; whether modern Hebrew is old or new… beats me.

            “in the older days, one would go to church”

            Attended by a local priest who tried to do most of the rituals in a Latin he barely understood anymore… yet I can still read Cervantes.

            I have the feeling that, because of the nature of the discipline, many linguists idealize written language, when in reality is in most cases nothing but a peculiar form of spoken language. Of course written language can and does affect spoken language but, unless most people is literate, not that much in fact.

            Granada was taught to Speak Castilian (Spanish) with one of the oldest standard grammars of living languages, created ex-professo for the assimilation of the conquered people, yet they still speak Andalusian and not ‘proper Castilian’ (mutually intelligible with some effort but not the same). TV (and surely radio and cinema before it) has had much more of an effect in establishing a standard accent and most common dictionary – but always up to a point: in spite of all an Andalusian great-grandparent will understand his/her great-grandson/daughter (and vice versa) much easier than I can.

            Because language is normally learned at home first of all.

          • Many linguists once thought that linguistic change in non-literate societies occurred at a regular pace, hence the field of “glottochronology.” Do I take it correctly that such ideas are now discredited, and that the pace of change is regarded as irregular and highly variable?  Also, is there something peculiar about Australian Aboriginal languages that makes them change so quickly? Aboriginal Australia was characterized by much long distance movement, spouse exchange, and ethnic fluidity, which might accelerate language change.

             Maju’s comments on the conservatism of Lithuanian is also intriguing, as I have often been puzzled by it. The Lithuanians also seem to have retained some very old cultural practices, such as honoring snakes — or so I have read.

    • Thanks, Maju. The Serbian nationalist views on the Albanian-Albanian connection is most interesting.  Do you now if any Balkan Albanians see any connections with Caucasian Albania? 

      • I know that they do not: it’s mere Serbian nationalist propaganda for people with low intellectual standards like the usual cannon fodder of all fascist organizations.

        As you may know, the origin of Albanians and specially Albanian language is the object of some debate and not fully clarified. Essentially there are two main schools: the Illyrianist (generally favored by Albanians) and the Daco-Thracianist, both of which are problematic (there are also hybrid theories for what I know: like Illyrian with Thracian influx, etc.)

        However, Serbian nationalism, specially the irrational maximalist kind that flourished around the conflicts of Krajina, Bosnia and Kosova, promoted the idea at times of Albanians (with whom they had, and still have, the conflict around Kosova) being late immigrants to the Balcans and being somehow related to Caucasian Albania. That hypothesis implied that Caucasian Albanians spoke an Indoeuropean language, the direct ancestor of Balcanic Albanian.

        It was just some propaganda to claim that Albanians had “stolen” the land to the Serbs. There is no substance to it other than the name. For the same “logic” they could have said they were from Scotland (Alba) or whatever.

        Incidentally the endonym of Albanian language is ‘Shqip’ in native albanian language, ‘Shqiperia’ being Albania (although the exonyms Alban- and Arvan- are very old and consolidated and Albanian speakers from Italy and Greece use them).

        • Very interesting comment, thank you!

          The crucial thing for these Serbian nationalists would be timing though: even if Albanians came from the Caucasus (or Scotland, as you say, very cutely!), Serbs too came from somewhere else (further north, generally speaking). For Albanians to have “stolen” Serbian land, they must have come there *later* than the Serbs. I wonder if the Serbian nationalists address this point at all…

          • badhistory

            if the serbs arent native to the balkans then show me some proof of where the native people went or the battles that took place to displace them.

          • Slavs are not native to the Balkans, but to an area further north. Serbs are mostly Slavs, although they have genetic admixture of various other people, including those who inhabited the area prior to the Slavs’ arrival (Illyrians, Greeks, etc.). Those groups were displaced further south as well as incorporated into the Slavic-speaking groups. Not sure why you are asking about battles: most population dispacement happens without specific battles. It’s a much more gradual and slow process, taking decades, centuries, sometimes millenia. What battles were fought to displace Dravidians from northern India? Finno-Ugric-speaking peoples from northern Russia? Khoisan-speaking groups from central Africa?

      • olympia

        They don’t. Their nationalism does not permit it.

  • David Erschler

    Udi & Albanian

    (Disclaimer: I haven’t ever worked on Udi myself.)

    Udis live in one village in Georgia as well (Zinobiani aka Oktomberi), on the other hand, a large number of Udis were expelled from Azerbaydzhan in late 80’s — early 90’s as Armenians. (Udis have Armenian names and surnames.) In particular, the majority of Udis from Vartashen were expelled. Vartashen was renamed by Azeris into Oghuz.

    Alexidze did not work alone: the involvement of Wolfgang Schulze and Jost Gippert was crucial (cf Aleksidze’s account of the discovery,

    The language significantly differs from modern Udi (and, moreover, modern Udi is probably a descendant of a closely related language variety, but not literally from the one of the palimpsest.)

    • Thank you for your informative comments, David! Have you had a chance to visit any of these Udi villages?

      I am glad you agree with my assessment that Ancient Albanian and Udi were quite different. It is easy to get carried away if you are a person deciphering the former…

      • David Erschler

        No, but some of my Moscow colleagues did (Yura Lander, Timur Maysak, and Dima Ganenkov)

        • I’ve met Yura Lander, I think, but never had a chance to talk to him about Udi.

    • Thanks again for the clarification and additional information — most useful. Do you know how many Udis were expelled in the late ’80s and early ’90s, and what became of them?

      • David Erschler

        Apparently they got dispersed in the countries of the former Soviet Union. I know about one village in Volgograd oblast’, Dubovy Ovrag, where Udis live compactly.

        I doubt that there are precise statistics on the number of Udis who left Azerbaydzhan. At any rate, the Russian 2002 census ( lists 3,721 Udis in Russia.

        • Thank you for the comment and the link to the Russian census — a very useful resource indeed.

  • Even if the gap between Udi and Albanian is fairly large, just having a relative makes a huge difference.  Classic Maya was worked on for a century before anyone seriously tried who knew the modern Maya languages, and then the walls came tumbling down in a single decade.  Linear B is Greek, and so we can read it; Linear A is not, and we have no clue.  With enough material and enough patience we have deciphered Sumerian, but rongo-rongo remains a complete mystery even though Pascuan, the language being written, still survives.  And as for the Harappan inscriptions, they are so deeply mysterious that there is not even agreement on whether they are writing.  (I find the arguments that they are not writing to be convincing, but certainly it is a minority view.)

  • In response to Maju’s comment above:

    1) Curious how you think that something is wrong if it doesn’t fit your preconceived notions. The best discoveries in science are in fact things that “simply couldn’t be”, but are!

    2) Australian aboriginal languages in question have had very little influence from English (and certainly no creolization). In fact, they barely have a handful of English loanwords to show for it.

    3) Re: Hebrew, even though it stopped being a spoken language, people continued to write in it, even in forms different from Biblical Hebrew (cf. e.g. Mishnaic Hebrew, which I had in mind in an earlier response)

    4) Your comment about the priest conducting mass is well-taken — for the Catholic world. World-wide however religious service has mostly been done in the local language, including most Orthodox Christian communities.

    5) Re: spoken vs. written language, once again you ascribe an amateurish view to professional linguists. The latter believe in the primacy of the spoken language as the object of study ( However, there is no denying that the availability of a written form, even if literacy is limited in the population, is an important factor affecting language change. To bring us back to the region under discussion, a Georgian or Armenian priest would read a religious text in Georgian/Armenian to the (mostly illiterate) congregation thus exposing them to a standardized form of the language.

  • In response to Martin’s comment above:

    1) The earlier assumption that the rate of change is constant have been pretty much proven wrong, so the classical “glottochronology” is not very fashionable any more. In recent years, however, people have been developing new and improved glottochronology, relying on methods from population genetics (esp. their statistical methods). Criucially, these methods
    calibrate points on the tree with known historical events and smooth the
    rates of change across these. As such, they no longer require the
    assumption of a constant rate of change. Moreover, these (and the classical glottochronology methods) work best over relatively long time periods, measured in thousands of years.

    2) Australian Aboriginal languages may be exceptional in their quick rate of change, or not. We just don’t know enough about how (quickly) languages of small tribal groups without writing change (after all, if they don’t have writing, our ability to know what they were like at any given time in the past is very limited, no?). You are absolutely correct about the cultural practices like long-distance marriages that may affect the rate of language change in Australia. Another reason is the extensive taboo systems, such that for example when a person dies, his name and all words that sound like it must be replaced with different sounding ones. To give an English-language parallel, imagine if a guy named Bill dies and you no longer can ask for a “bill” in a restaurant… (nor can a bird have a “bill”), or if a gal named Pat dies and you can no longer “pat” your “pet”…

    • “… glottochronology, relying on methods from population genetics (esp. their statistical methods)”

      Age estimation in population genetics is as
      of today mere pseudoscience (or when done more respectably an educated
      guess). Not only they have the bad custom of undercalibrating key
      references like the Pan-Homo divergence point but what you claim of
      calibrating points across the tree (I presume in relation to real historical or prehistorical events) is seldom done if at all.

      I sincerely hope that “new glottochronology” is more than that because molecular-clock-o-logy is, at the best, a good hunch and at the worst a bad hunch pretending to be “science” without any grounds (pseudoscience therefore).

      “… these (and the classical glottochronology methods) work best over relatively long time periods, measured in thousands of years”.

      Sounds like hiding in the darkness of prehistory. I don’t say it’s wrong but sounds like impossible to prove correct: how can I accept that a chronological estimation method which is unable to correctly estimate something known (like what happened in the last few centuries) can estimate correctly something unknown?

      When C-14 age estimation methods were discovered and presented to the world, their accuracy was once and again tested on objects of known age. Only then C-14 became accepted (even if it’d be refined later anyhow). Today any academic hunch with some statistical methodology is considered “scientific truth”, what is a shame! And a threat to science as such methodology of systematic doubt.

      “Another reason is the extensive taboo systems, such that for example
      when a person dies, his name and all words that sound like it must be
      replaced with different sounding ones”.

      That’s pretty impressive and seems to explain on its own the extremely fast rate of change.

      However this renders the Australian example useless for the rest of the Universe.

      • Re: glottochronology, different scientific techniques work on different temporal range, just like a microscope and a telescope would be use for seeing objects of different sizes and at different distances. Just like a telescope is not useful for searching for a needle in a haystack, glottochronology is not good for dating events that are too recent.

        Re: taboos and the rate of language change, taboo systems are actually quite common in tribal societies around the world. Hence, the Australian example is not “useless for the rest of the Universe”, as you claim.

  • Pingback: [LINK] “Udi and Ancient Albanians” « A Bit More Detail()

  • Pingback: Loanwords in Armenian | Etymology | Languages Of The World()

  • Arentma

    “At roughly the same time, the land of Urartu seems to have been linguistically transformed by the spread of proto-Armenians from the west, a people perhaps linked with the ancient Phrygians who spoke a language in an outlying branch of the Indo-European family. In the twentieth century, Armenian nationalists began to glorify ancient Urartu as the deep font of Armenian culture. In doing so, they sought to highlight the antiquity of their claims to territory in what is now eastern Turkey. Without endorsing such political claims, it is only fair to acknowledge a close historical connection between Urartu and Armenia.”

    Thanks for the comprehensive article. The passage above is quite well-balanced in regards to the linguistic stratification of the Armenian language (even concerning the over-nationalization of Urartu), but one of the unresolved components of the “Armenification” formula merits at least a review, if not a general reconsideration: Just how far west the linguistic carriers of (Proto-)Armenian are localized at the onset the post-Urartian movement (that is, if they did not already constitute a sizable percentage of the empire’s population) is central to the unbiased, academic understanding of Armenian ethnogenesis. The further identification of Hayasa-Azza, the ethno-tribal confederation mentioned in Hittite sources immediately to the northeast of the Anatolian kingdom, is particularly pertinent.

  • Beno

    I have a question that intrigues me, and I wonder if you can answer me: the Udi Church is recognized as an orthodox church? If yes, would be an autocephalous orthodox church or be part of the Russian Church? If not, what is the affiliation of this christian community?

    Thanks for your attention, and I apologize for commenting on an old post..

    • Thank you for your interesting question, Beno. As far as I understand, the Udi church used to be an independent autocephalous church, but then fell under the religious jurisdiction of the Armenian Apostolic Church. It is thus not part of the Russian Church and technically not even Orthodox.

  • χατιρι αζερικα

    udis were never expleled from Azerbaijan and they still live in different areas, in Gabala (Azerbaijan) the capital of ancient Albania, which actually was comprised of many people who spoke in 26 different languages, you can check Mouises Kalankatlu, one of the earlierst historians writing about the region.

  • χατιρι αζερικα

    the Albanian church and all resources in Albanian language were usurpated, falsified or annulled by Armenians after tsarist regime was established in those regions and the armenians from Turkey, Iran were located in the Caucasus in XIX century, there was a monument in Sumgayit city dedicated to their arrivall (with the mark 150 years) which was abolished as well lateron