I posted briefly about the many languages of the Caucasus here; I was very pleased to discover that GeoCurrents has been creating a more accurate map of them than has been available, as described in this post at their site:

Drawing on previously available ethnic and linguistic maps, supplemented by demographic data from other sources, we were able to create two linguistic maps: one representing the whole Caucasus area and the other zooming in on the particularly linguistically diverse region of Dagestan. Our first task was an accurate representation of the spatial distribution of various groups, unlike what is found in previously available maps, which often over-represent or under-represent the extent of linguistic groups. We have used the most recent census data available to capture the wholesale migrations, episodes of ethnic cleansing, and population exchanges that have changed the situation on the ground. Careful mapping of smaller linguistic groups, especially in Dagestan, has proved particularly instructive, as it allowed us to represent visually the correlation of language and topography, something that has not been done before. …. Finally, a careful use of the color scheme allowed us to demonstrate the family relatedness of the various languages spoken in this region, known justifiably as “the mountain of tongues”.

They welcome “comments and corrections from informed readers, especially those who live in the Caucasus or have done fieldwork there.”

Update (Dec. 2022). There is a newer (but, annoyingly, undated) version of the GeoCurrents post here, with revised maps.


  1. megazver says

    Links to illegible thumbnails.

  2. The illegible thumbnails look similar to the high resolution maps, but they have hatched areas where the large images have white or monochrome. I wonder which version is the latest.

  3. Sorry about that; I copied the links in their post and didn’t even check them. But of course you can get to the maps at the main link.

  4. Siganus Sutor says

    Hmmm, there must be something wrong. I don’t see any Dravidian language anywhere on this map. (Unless the great Batsbi, isolated as it is high in the mountains of Georgia, is one of them?) It is urgent that John Emerson be called to have a closer look at all this.

  5. A lecture on Dagestanian dialects read by Nina Sumbatova in the Polytechnical Museum in Moscow (in Russian)

  6. dearieme says

    Meanwhile, further east: Burushaski.

  7. These blog entries about Burushaski I think are informative

  8. Oh, that’s where the most recent Burushaski flare-up registered! Just a few hours earlier, I emailed Language about this news release, with its oops-quality juxtaposition of Indo-European vs. Indo-Iranian languages.
    The Burusho used to fascinate the geneticists of human diversity, ever since the other language isolate, the Basque, was shown to be associated with a strong, ancient genetic footprint in human DNA in a wide area of Europe. But the Burusho DNA didn’t yield any similar clues about the prehistory of the human population in the region. The language is very unique but the genetic makeup is more or less typical for the area.
    BTW in this post, I’m trying to switch from a semi-anonymous username to my regular human name. Is it an OK thing on languagehat, to “change names”? Or should I continue to stick to the old nom-de-net “MOCKBA” for contunuity?

  9. Sure, you can call yourself whatever you want (just ask AJP, and John Emerson used to be Zizka).

  10. just ask AJP
    LOL but there is a good deal of continuity in all those AJP names 🙂

  11. marie-lucie says

    John Emerson used to be Zizka
    I never suspected! I wondered what had happened to Zizka.

  12. Trond Engen says

    If you see a comment you really like, the safest thing is to assume it’s me using another name.

  13. Trond Engen says

    Dmitriy Pruss: I’m trying to switch from a semi-anonymous username to my regular human name.
    And adding a link to an impressive list of publications in a field I hadn’t imagined. Well done!

  14. Trond Engen says

    It follows that they were written by me using your name. Hope you don’t mind.

  15. If you see a comment you really like, and it isn’t signed “John Cowan”, the safest thing is to assume it isn’t me using another name. (Well, except “Eoghan Mac Eoghain”, which is the same name, only in Irish.)
    Exchange between Franklin Knight Lane, 1902 candidate for governor of California, and William Randolph Hearst, newspaper magnate. After Hearst’s papers had slanted their news coverage against Lane; Hearst denied that this was his personal policy:

    “Mr. Lane, if you ever wish anything that I can do, all you will have to do will be to send me a telegram asking and it will be done.”

    “Mr. Hearst, if you ever get a telegram from me asking you to do anything, you can put the telegram down as a forgery.”

  16. I don’t actually think the press release is particularly problematic: it should just read “Indo-European in origin, but not Indo-Iranian” instead of “Indo-European in origin, not Indo-Iranian.” This is confirmed by various other online references to Čašule’s theory.
    Here’s a brief negative review by Ed Vajda of Čašule’s 2010 book, and an earlier review, also negative, by Bengtson and Blažek of his 2003 JIEL article.

  17. Thanks. Sounds like Čašule’s arguments aren’t very convincing.

  18. Trond Engen says

    Following Steven Lubman’s link, Asya Pereltsvaig says:

    The one important generalization to draw from this is that grammatical features (or features of the sound system) by themselves cannot be used as argument for (or against) language relatedness. What needs to be found is convincing evidence of lexical cognates. In the case of Burushaski, such cognates are absent.

    True, of course, but I’d say that for a highly agglutinative language like Burushaski, the real test is the morphology. Without cognate morphemes and a plausible explanation of its development, even a large number of regular cognates would at the most amount to an old adstratum.

  19. marie-lucie says

    Thanks for the links, JC. Someone else had given the second one in a facebook comment, and that’s how I was able to read it. It is quite evident from the Burushaski data that the language is extremely different from any Indo-European language in grammar and basic (non-borrowed) vocabulary, but it seems to be much closer to some Caucasian languages (Ed Vajda agrees with Bengtson & Blažek on all those points). Those criticisms relate to earlier works by Čašule, but judging from the information in the more recent news item it does not seem that Čašule has changed much in his methods.
    I agree with Trond about Asya Pereltsvaig’s comment (she is a typologist, not a historical linguist). Features of grammar (such as the existence of gender, tense, etc, or the basic word order) and phonology (such as type and number of vowels and consonants), and many others, can develop independently within widely separated languages, and they can also be shared by unrelated languages spoken in the same geographical area by many bilinguals. Related languages have many related grammatical morphemes as well as related words. Unrelated languages may express the same grammatical meanings but with quite different morphemes. For instance, the other day there was a discussion about how to pluralize Russian drug when used in an English sentence: English speakers ignorant of Russian may happily borrow drug but will naturally add to it the English plural morpheme -s, just as those ignorant of Latin will naturally pluralize cactus as cactuses rather than wait to learn the Latin cacti (in both cases, the foreign form is only one in several other plural forms anyway).
    Attention to grammatical morphemes is the missing element in much current historical linguistics, which tends to rely on lexical items (basically, vocabulary words). In the early days of historical linguistics, close attention was paid to morphological structure and individual morphemes, which played a crucial role in establishing the genetic relationship of the Indo-European languages. Once that relationship was basically understood, linguists turned to the reconstruction of the common ancestor “Proto-Indo-European”, and concentrated on the lexical items since vocabulary words are immensely more numerous than grammatical morphemes. As a result, many people (like Asya) think that lexical cognates provide the most crucial evidence. But it is quite possible for two languages to have large numbers of similar words linked by regular sound correspondences without being closely related: French and English are an obvious example, since the existence of numerous similar words is due to a fact of history which brought French speakers to England and caused English speakers to borrow many French words, which have evolved phonologically alongside the native English words while French words in France evolved in their own way. It is true that the two languages are related, but much more distantly, and they are not each other’s closest relatives in spite of sharing a lot of similar vocabulary: English is most closely related to Dutch and German, French to Spanish and Italian (among others). This conclusion arises from even a casual look at verb formation and a few other grammatical features and morphemes (eg the articles) in those two sets of languages. In the Burushaski case, the Indo-Iranian hypothesis first arose because of numerous words borrowed into Burushaski from neighbouring Indo-Iranian languages, a subset of the larger Indo-European group. When those borrowings are excluded, the rest of the vocabulary, and especially the grammatical structure and individual morphemes, are very different from their Indo-European counterparts.

  20. Trond Engen says

    marie-lucie: I agree with Trond […] Attention to grammatical morphemes is the missing element in much current historical linguistics, which tends to rely on lexical items (basically, vocabulary words). […] As a result, many people (like Asya) think that lexical cognates provide the most crucial evidence.
    Thank you. May I go further out on the limb (a grave sin for a structural engineer) and say that I found it somewhat odd that Edward Vajda didn’t make that point in his review? To me his slot-by-slot and morpheme-by-morpheme comparison of Yeniseian and Na-Dene paradigms seems much more convincing than any list of cognate words could possibly do. But I remember him saying in an interview that he needed to put much more work into establishing cognates before his hypothesis is entirely convincing, so maybe he shares the modern misconception. Or at least thinks most of his peers do.

  21. The claim that the Yenisean and Dine language families is an extraordinary one (because of the physical distance and resulting time-depth) and requires extraordinary evidence, which Vajda is eager to discover. The claim that Burushaski is an IE language is still more extraordinary, and would require evidence beyond a reasonable doubt, which Čašule plainly has not got.

  22. marie-lucie says

    JC: good comparison (or tather contrast).
    The current geographical gap between Yeniseian and Dene speakers is indeed very large, but “Yeniseian” refers to the historical location of the known members of this family, the “cradle” of which may have been elsewhere. The still surviving Yeniseian language (Ket) is only one of a number of related languages which are known to have existed. Some of those, now unfortunately extinct, were investigated by Russian scholars while they were still spoken, and others are known more fragmentarily from earlier Russian sources, so that there is a fair amount of information on the family. Also, Yeniseian names of some Siberian rivers outside of the Yenisei valley (river names are famously conservative) show that Yeniseian languages were once spoken over a much wider area. Given the nomadic way of life of many Siberian tribes as well as of Alaskan and Northern Canadian Dene people, and the migrations which brought some Dene people southwards into Oregon, California and Southwest US states, a migration or series of migrations through Northern Siberia over centuries and even millennia is not out of the question.
    Here I must say that I disagree with the “one migration” theory bringing humans on foot from Asia to the Americas before the opening of the Bering Strait – I think there were several migrations, bringing a variety of languages and cultures into the Americas, many of them by sea (whether voluntarily or by accident). In the Dene-Yeniseian case, the numerous resemblances noticed earlier by some scholars, and consolidated and expanded by Vajda, would be extremely unlikely to have survived for 12,000 years or so if the split between the Siberian and American branches had occurred that far back in time. No known language family is that ancient.
    About the lexical cognates, Vajda has indeed been working on those, and it was thanks to this part of his work that he was able to convince some skeptical linguists at the conference at which he announced his conclusions (in 2008). His work also validated the earlier work of linguists who had noticed resemblances between Yeniseian and Dene, without having the means of substantiating them (for decades Russian scholars were not in a position to come to Alaska or Canada to study Dene languages in situ, and similarly it was only after the breakup of the Soviet Union that Vajda – a fluent speaker and teacher of Russian – was able to travel to Ket territory).
    On the other hand, with Burushaski Čašule reintroduces a hypothesis which is one of many that have been proposed and discarded over the years. Yeniseian and Dene languages are not very well-known except to a few specialists (and very few of those are or were competent in both families), and while Burushaski is also little-known, Indo-European languages are certainly not lacking in documentation, and comparative/historical work on them has been involving large numbers of linguists for over two centuries. Many persons who have taken an interest in Burushaski are or were quite familiar with Indo-European studies and issues. The methods originally developed for Indo-European languages have proven their usefulness all over the world, in establishing language families and reconstructing their ancestor, or conversely in discarding hypotheses which could not be substantiated. When Hittite and Tocharian writings were first discovered and deciphered, it took a very short time to identify the languages as Indo-European, but Burushaski has defied all attempts at linking it to any other known languages: every such hypothesis, when carefully examined against the known facts, has been discarded.. Čašule thinks he can rehabilitate the discredited hypothesis of Burushaski-Indo-European relationship by using a new method of his own invention to show how “languages blend together”, rather than evolve according to certain well-known patterns. “Blended” languages are not unknown: pidgins, creoles, and a few other types s have been extensively studied, including the circumstances under which they arise. Čašule’s alleged method does not build upon other scholars’ work but discredits it. In this respect it is more reminiscent of Greenberg and Ruhlen’s elucubrations than of the painstaking work of true historical linguists.

  23. m-l: I agree on all points except the question of time-depth. The fastest possible separation would be for the Dine-Yeniseian homeland to be at the halfway point exactly between the Dine and Ket homelands, and for the two groups to travel east and west in a straight line. The best figure I can find for hunter-gatherer migration rates is 5 km/year, so it will take a loooong time for both groups to make the trip, even though each group only has to travel half the distance. Straight-line migration directly from Ketland to Dineland (or vice versa) would take twice as long, and zig-zagging (which is undoubtedly what happened) would make progress even slower.
    There was another D-Y conference last March, but the only thing published so far is a bunch of video lectures on YouTube. Vajda’s is called “Geography, Demography, and Time-Depth: Explaining How The Dene-Yeniseian Hypothesis Is Possible”. I hope to get a chance to watch it early next month.

Speak Your Mind