May 24, 2011

The reality of the Altaic language family

Personally I'm not surprised by this; my own look at genomic data has identified an "Altaic" component which peaks at the Turkic Yakut and Tungusic Evenk, and is shared by every Turkic, Mongolic, and Tungusic population available to me. The same component also occurs to some extent among all the Japanese (5) and Korean (4) members of the Dodecad Project, while it is lacking in all the Chinese ones (8).


Of particular interest is the degree of CCM between Indo-European and Semitic languages (Tables 2 and 3). In many of the most geographically distant languages these are less than 10; by comparison, among Semitic languages the are all greater than 20. This seems to be quite in agreement with the idea that Semitic is a Bronze Age language family, Indo-European a Neolithic one.

This impression is strengthened by the fact that CCM between reconstructed proto-languages (e.g. Proto-Iranian and Proto-Slavic = 20) are much higher. Since these proto-languages are a few thousand years closer to the root of PIE than present-day languages, and differences between them are similar to those of Semitic languages, the notion that PIE is a few thousand years older than Proto-Semitic seems quite consistent with the evidence.

Journal of Language Relationship • Вопросы языкового родства • 3 (2010) • Pp. 117–126 • © Turchin P., Peiros I., Gell-Mann M., 2010

Analyzing genetic connections between languages by matching consonant classes

Peter Turchin (University of Connecticut)
Ilia Peiros (Santa Fe Institute)
Murray Gell-Mann (Santa Fe Institute)

The idea that the Turkic, Mongolian, Tungusic, Korean, and Japanese languages are genetically related (the “Altaic hypothesis”) remains controversial within the linguistic community. In an effort to resolve such controversies, we propose a simple approach to analyzing genetic connections between languages. The Consonant Class Matching (CCM) method uses strict phonological identification and permits no changes in meanings. This allows us to estimate the probability that the observed similarities between a pair (or more) of languages occurred by chance alone. The CCM procedure yields reliable statistical inferences about historical connections between languages: it classifies languages correctly for well-known families (Indo-European and Semitic) and does not appear to yield false positives. The quantitative patterns of similarity that we document for languages within the Altaic family are similar to those in the non-controversial Indo-European family. Thus, if the Indo-European family is accepted as real, the same conclusion should also apply to the Altaic family.

Link (pdf)

29 comments:

  1. Personally I'm not surprised by this; my own look at genomic data has identified an "Altaic" component which peaks at the Turkic Yakut and Tungusic Evenk, and is shared by every Turkic, Mongolic, and Tungusic population available to me. The same component also occurs to some extent among all the Japanese (5) and Korean (4) members of the Dodecad Project, while it is lacking in all the Chinese ones (8).

    According to the results of the paper, the Eskimo-Aleut language family is related to the Altaic language family. This further confirms that the Altaic language family originated in East Siberia. Then the question to be asked is where the five main branches of the Altaic language family (Turkic, Mongolic, Tungusic, Korean and Japonic) originated, and all the available evidence so far points to anywhere from the northern regions of East Siberia to the northern (especially northeastern) regions of China for all of the five main branches.

    ReplyDelete
  2. Murray Gell-Mann? Really?

    Anyway, simple but clearly defined tools like this one have their place in answering important questions. In another life (meaning having more time on my hands) I definitely would have pursued this type of avenue.

    My only worry is that reconstructed proto-languages are not "independent." They are often built just for the purpose of demonstrating/highlighting relations. As such, (i) it should not be a surprise that they do so well, here, and (ii) this study needs a more careful evaluation of this type of effect, and for that and other reasons should be broadened to other linguistic parameters.

    ReplyDelete
  3. So Japanese is Altaic... even though we just saw that it is supposed to have come to Japan recently and probably with thecrice farmers.

    ReplyDelete
  4. "all the available evidence so far points to anywhere from the northern regions of East Siberia to the northern (especially northeastern) regions of China for all of the five main branches".

    I think the evidence suggests that Tungusic's movement north into the Lower Lena valley, for example, is relatively recent. So that would place Altaic's origin somewhere in the mountains of Northern Mongolia, the Upper Amur.

    The idea that Japanese and Korean are part of that group has a long history, although vehemently denied by some.

    ReplyDelete
  5. This comment has been removed by the author.

    ReplyDelete
  6. So that would place Altaic's origin somewhere in the mountains of Northern Mongolia, the Upper Amur.

    Actually the likely homelands of the five main branches of the Altaic language family (=the territories from the northern regions of East Siberia to the northern - especially northeastern - regions of what is now China) are also the likely homelands of the Altaic language family itself. So what is now Northern Mongolia is a likely homeland of the Altaic language family too.

    ReplyDelete
  7. "Murray Gell-Mann? Really?"

    Yes, he's been passionate about long-range historical linguistics as a hobby for quite some time now. I met him at the Greenberg Memorial Conference in 2003 or 2004. He was affiliated with the Santa Fe Institute already then.

    With all these papers offering new computerized, statistical proofs for old ideas, I need to see responses from the specialists in those languages who are anti-Altaicists. If they snidely refer to the voluminous Etymological Dictionary of the Altaic languages as a "dictionary for a non-existing family" (Georg 2004), I doubt they are going to be convinced by this skin deep approach.

    "So Japanese is Altaic... even though we just saw that it is supposed to have come to Japan recently and probably with thecrice farmers."

    Good point. Very good point, in fact. If I were choosing, I would ignore the Bayesian paper.

    "the Eskimo-Aleut language family is related to the Altaic language family. This further confirms that the Altaic language family originated in East Siberia."

    The Eskimo-Aleut family is also being reliably (by long-range comparison standards) connected to Uralic but using other than consonant matching method data (see Seefloth's work on pronominal paradigms). This squares well with the origin of Uralic in East Siberia but also makes the picture a bit messier as Uralic and Altaic are currently believed to share only areal similarities.

    ReplyDelete
  8. "even though we just saw that it is supposed to have come to Japan recently and probably with thecrice farmers".

    Surely that in no way prevents it belonging to 'Altaic'.

    ReplyDelete
  9. "even though we just saw that it is supposed to have come to Japan recently and probably with thecrice farmers".

    Surely that in no way prevents it belonging to 'Altaic'.


    I totally agree with you, Terry.

    ReplyDelete
  10. "Surely that in no way prevents it belonging to 'Altaic'."

    Ya, your right. It just surprises me that the rice farmers coming to Japan recently would have been Altaic speakers.

    ReplyDelete
  11. "Surely that in no way prevents it belonging to 'Altaic'."

    According to glottochronology, Turco-Mongolic languages diverged 6000 years ago. The Korean-Japanese node is less secure than the Turco-Mongolic one but, provided that all the cognates have been correctly identified, then it's some 5000 years old. The onset of Yayoi is 300 BC and that's where the Bayesian method put proto-Japonic. We have a gap of 2500 years to fill. It IS a problem.

    ReplyDelete
  12. I don't understand why some people think rice farmers cannot be Altaic-speaking.

    If you assume Korean to be Altaic then there was at least one altaic speaking rice farmers for more than 1000 years before Yayoi migration took place.

    There clearly were Japonic speaking people in Korea before the Yayoi migration. This is undisputed. Whether they were Koguryonic is what is being disputed.

    The linguistic landscape of ancient Korea is complex. It is explicitly stated in historical texts even though they don't clarify what those languages were.

    ReplyDelete
  13. "According to glottochronology, Turco-Mongolic languages diverged 6000 years ago. The Korean-Japanese node is less secure than the Turco-Mongolic one but, provided that all the cognates have been correctly identified, then it's some 5000 years old. The onset of Yayoi is 300 BC and that's where the Bayesian method put proto-Japonic. We have a gap of 2500 years to fill. It IS a problem."

    I posted an example of a three-way comparison of English, Korean, and Japanese "basic vocabulary" according to the Swadesh list on this thread yesterday, but it seems to have been censored by Mr. Pontikos. To summarize, I cannot imagine how one could use glottochronology to date the hypothetical split between the Japanese and Korean languages since there are no more similarities between the Japanese and Korean Swadesh list-based vocabulary lists than may be expected to occur due to random coincidence.

    ReplyDelete
  14. "If you assume Korean to be Altaic then there was at least one altaic speaking rice farmers for more than 1000 years before Yayoi migration took place."

    Japonic is not a dialect of Korean.

    "There clearly were Japonic speaking people in Korea before the Yayoi migration. This is undisputed."

    What's the evidence for that?

    ReplyDelete
  15. JCA or Ebizur erred mightily in his overzealous attempt to separate Korean and Japanese.

    Read Alexander Vovin's "Korea-Japonica: A Re-evaluation of a Common Genetic Origin" 2008.

    While his conclusion is also to doubt genetic(linguistic sense) relation between Korean and Japanese, his examples show abundance of similarity even in the basic vocabulary list.
    He successfully attributed most of these to the nearly unidirectional influence of Korean on Japanese from the late 3rd century to the early 8th century.

    JCA's claim may sound professional as it is what is claimed by the more careful linguists in this field but his methodology is amateurish.

    ReplyDelete
  16. @Ebizur

    "I posted an example of a three-way comparison of English, Korean, and Japanese "basic vocabulary" according to the Swadesh list on this thread yesterday, but it seems to have been censored by Mr. Pontikos. To summarize, I cannot imagine how one could use glottochronology to date the hypothetical split between the Japanese and Korean languages since there are no more similarities between the Japanese and Korean Swadesh list-based vocabulary lists than may be expected to occur due to random coincidence."

    This is exactly what linguists that argue against the Altaic connection of Japonic and Korean point out. I'm not a specialist in these languages, so I can't say, but I agree this is a valid criticism. BTW, my dates for divergences within Altaic come from the Russian version of wiki, which apparently got it from Dybo/Mudrak's Etymological Dictionary of Altaic languages.

    ReplyDelete
  17. The reference to Korean in the abstract when Korean is not one of the languages examined in the study verges on deceptive.

    It is notable that the closest Altaic match to Japanese is proto-Turkic and that other matches to Japanese seem to be mediated through it.

    The basic problem with the methodology is the assumption that a quite crude and judgment prone analysis can itself say very much, as opposed to a more multi-faceted, historically and archaeologically and genetically grounded one.

    Another important observation to make about Japanese is that more than 50% of its lexicon consists of words borrowed from Chinese and other non-Altaic languages in the historically documented era, with Chinese also serving as the dominant source for its written language (with most of the rest being local innovation). Regardless of the extent to which an Altaic language was the linguistically genetic ancestor of Japanese in 300 BCE give or take(and the proto-language was more similar grammatically and phonetically to Altaic as well as lexically than modern Japanese), modern Japanese is arguably more influenced by Chinese than any other language.

    Most professional accounts describe Korean as the language of one of the pre-unification kingdoms of Korea and Japanese as the probably descendant of another language of the pre-unification kingdoms of Korea, but there is contradictory professional opinion regarding which of the pre-unification kingdom's languages gave rise to which of the two modern languages, an undertaking further complicated by the shifting boundaries of the kingdoms in the period from which we derive evidence.

    ReplyDelete
  18. Chinese loan words in general are meaningless in historical linguistcs. It is like English in modern world.

    If you find English loan words in Russian it means virtually nothing. If you find Swahilli loan words on the other hand it may be interesting.

    Also Chinese loan words usually date after the introduction of Buddhism both in Korean and Japanese. It may be of cultural and historical interest but not too much for Korean and Japanese comparative linguistics.

    Actually its best usage is in relation to Old and Middle Chinese linguistics.

    ReplyDelete
  19. Due to Dienekes' continual censorship, I will not reply to any comment on this blog in the future. I have no interest in participating in a debate in which I am not allowed to present substantial evidence for my position, while an opponent is given free rein to make unsupportable claims and insidious ad hominem arguments that border on libel.

    ReplyDelete
  20. Due to Dienekes' continual censorship

    You triple posted.

    ReplyDelete
  21. Actually JCA has shown that he knows nothing about Korean. Check zetaboard run by Ren.

    How can he make a comparison with a language he knows nothing about?
    I was just stating the obvious.
    If you want to talk about Korean you need to know something about the language even if you childishly loathe the people who speak that language. At least minimally.

    He once conjectured that "nan" (which is a contraction of "na neun") derives from Chinese because of the ending n. When someone pointed out that it is a contraction that is regularly permitted with only a handful of exceptions he went berserk.

    I am not here to diss him but just wanted to show what his "expertise" in Korean is like.

    ReplyDelete
  22. Dienekes said,

    "You triple posted."

    I presented a single reply in three pieces in order to avoid Blogger's limit on the number of characters in a post. I think most would consider that to be different from posting the same content multiple times à la Onur.

    Hector said,

    "He once conjectured that "nan" (which is a contraction of "na neun") derives from Chinese because of the ending n. When someone pointed out that it is a contraction that is regularly permitted with only a handful of exceptions he went berserk."

    I regularly communicate with friends of mine in Korean, and I have researched many documents that are relevant to the reconstruction of earlier forms of Korean etyma, such as Gyerim Yusa (鷄林類事). I probably have a better grasp of the historical development of the Korean language than you do. If not, please present some of your own knowledge on the subject. All you have done up to this point is to spew unwarranted insults.

    As for my "conjecture" regarding the precedence of the short forms of pronouns in the Korean declensional paradigm, if you were really as erudite as you pretend to be, you should have taken note of my reference to Kim Bang-han's 한구어의 계통 "The Genealogy of the Korean Language" (1983).

    ReplyDelete
  23. LOL JCA, I have the book. It is "한국어의 계통", not "한구어...". It just shows how little you know about Korean.
    Kim says nothing of that sort. I seriously doubt you understood any of what he said given your poor command of the language even if you read a translated version.
    The main issue between Korean and Japanese is not chance-resemblance. It is way beyond that point. The issue is whether the similarity arose from contact or from common descent.

    Your conjecture... that was really seriously comical. Such a contraction happens just about anywhere. It can even combine with a regular noun. I am just laughing laughing laughing so hard.

    You somehow sensed that Koreans did not want to be associated with Chinese. That is why you were pushing that point. If you thought that Koreans wanted to be Chinese you would be pushing for the opposite case.

    ReplyDelete
  24. This comment has been removed by the author.

    ReplyDelete
  25. This is getting ridiculous.
    Someone who cannot distinguish "한국어" and "한구어" wants to keep talking about Korean. Unbelievable.

    If Romanized, it is "Hangugeo" and "Hangu-eo" respectively. You printed the title in Korean even though there was no need to, apparently to show off how much you know about the language.
    That failed comically and hilariously.

    You just have no credibility whatsoever. Are you really a linguistics major as you said? I just don't buy it. Not just with Korean, you are utterly unfamiliar with how comaparative linguistics is done.

    ReplyDelete
  26. Hector, your attemps to defame me by depending on others' ignorance of the Korean writing system are bound to fail in the end. The vocabulary included in the Swadesh list is called "basic vocabulary" for a reason, and Modern Korean basic vocabulary has changed only slightly from that of Middle Korean documents that I have used to create my Swadesh list of Korean vocabulary for comparison with Japonic. Do you really think it is so difficult for an English speaker to confirm that the Korean word for "long" is /kir-/ (or gil(-da) according to the currently popular system of romanization), the Korean word for "small" is /cak-/ (or jag(-da)), and the Korean word for "rain" is /pi/ (or bi)? These are not esoteric concepts, and Korean is not an obscure, undocumented language.

    ReplyDelete
  27. Well, Hector, at least that minor boo-boo (that a normal adult would skim over during reading rather than making such a fuss) shows that I have been typing Korean text on my own keyboard rather than copying and pasting. If you want to make an ass of yourself over a miniscule typo (which I only made because of Blogger's ridiculously narrow comment box and vexing auto-scroll function, as you should know if you have read my posts on Quetzalcoatl, but instead you have chosen to ignore the fact that I have quoted and discussed long passages from Kim Bang-han's book on that forum because it suits your present puerile rant), then I will treat you with an equally nitpicky and inimical attitude. Since you claim to possess the book in question, you should know that its title is not actually printed as 한국어의 계통. If you cannot reply with the exact title of that book, I will consider it to be an invalidation of your authority to discuss Japanese or Chinese historical linguistics. "남에게 대접을 받고자 하는 대로 남을 대접하라."

    ReplyDelete
  28. LOL someone is clearly riled.

    Typos can happen but the real comedy was his not recognizing it even when it was pointed out. And then he proceeded to make hilarious accusations based on that.

    But the real highlight was his deleting the post where he made an ass out of himself. This guy is pure sleaze.

    You got to see his deleted post to see what I mean. It was really funny, hysterically funny.

    ReplyDelete

Stay on topic. Be polite. Use facts and arguments. Be Brief. Do not post back to back comments in the same thread, unless you absolutely have to. Don't quote excessively. Google before you ask.