If you’ve scrolled down to the comment section of the video that goes along with today’s blog post, it won’t be hard for you to realize that the Altaic language family is a very controversial topic in linguistics.
I apologize in advance for the offensive language from some of our commenters.
So, what makes the Altaic languages so controversial?
A Note on Language Families
Well, I’ve spoken about them briefly before in the Languages of Russia video, and in short, the Altaic Languages are a proposed language family. Currently, it is estimated that there are about 150 language families on Earth, not including language isolates, but only 18 of these language families contain at least 1% of the over 7,000 known languages on Earth.
To summarize, the largest language family is the Niger-Congo languages followed by the Austronesian languages which includes languages like Malay and Tagalog. The Indo-European languages come in a number 5 despite the fact that they are the most spoken with almost 3 billion combined speakers, and the Afro-Asiatic languages, which includes languages like Arabic and Hebrew, are at number 7.
The Altaic Language Family
What does our proposed theory have to do with this list? Well, if proven to be true, the Altaic language family may be included to such a list, albeit with not so many languages, but with a lot of speakers.
As for the theory itself, the initial seeds for the Altaic language family theory was in 1844 when a Finnish philologist by the name of Matthias Castrén was studying the Uralic language family. Matthias proposed that the Uralic languages, including Finnish, Hungarian, and Estonian, were, in fact, related to other languages like Turkish and Mongolian.
Over the years, this initial theory changed: the Uralic languages got taken out of the equation, but Japanese, Korean and the Tungusic languages got added, and by the 1960’s, the proposed Altaic languages attempted to unite large sections of Asia. The theory encompassed about 75 individual languages (more if one included the Uralic languages), most or all of which were said to be related because they shared common features such as a subject-object-verb word order, a lack of grammatical gender, vowel harmony, and highly-agglutinative grammars.
Further research was done throughout the 1980’s and 1990’s, and the Altaic language family reached quasi-acceptance around 1991 when a researcher named Sergei Starostin used 110 of the most common words across all languages (called the Swadesh-Yakhontov list) to compare the languages of the proposed Altaic language family. He found the following percentages of lexical similarity (or percentage of words with common roots between two languages):
- Turkic–Mongolic: 20%
- Turkic–Tungusic: 18%
- Turkic–Korean: 17%
- Mongolic–Tungusic: 22%
- Mongolic–Korean: 16%
- Tungusic–Korean: 21%
Further, in 2003, Starostin and two other linguists published “An Etymological Dictionary of the Altaic Languages” which described some 2,600 cognates between the proposed Altaic languages, and this time, Japanese included. It seemed a new language family had been born.
Why is This Theory Controversial?
“So, Michael,” you’re probably saying. “With all this supposed grammatical and lexical evidence for the Altaic language family, why is this topic so controversial?”
Firstly, critics of this theory point to what is called “language contact” to explain the similarities between these languages. This means that because, for example, the Turkic languages and the Mongolic languages are spoken relatively close to one another, they are more susceptible to borrowing each other’s words. Besides, just because two languages have cognates or shared grammatical features, it doesn’t mean they are necessarily related.
To prove this, we can look at English and Zulu. Both of these languages have a subject-verb-object word order, and English has borrowed a number of Zulu words such as “impala” and “vuvuzela”. However, we know that these two languages are not linguistically related.
Related: Sanskrit & Armenian: Mothers of ALL Indo-European Languages?
The Death of the Altaic Language Theory
The biggest issue with the Altaic languages, though, goes back to a basic concept in linguistics.
Through historical linguistics, we can trace a language’s evolution through time, and by combining historical records and educated guesses, we can theorize what languages sounded like a long time ago, and even see how older languages split and changed into the modern languages we see today.
For example, we can trace modern Turkish back to a common ancestor shared with Azerbaijani, and that common ancestor had a common ancestor with the Turkmen language. At one point, this ancestor called Oghuz, came from Proto-Turkic, the ancestor of all the Turkic languages. Through this analysis, we can see how the languages have become more dissimilar over time, their mutual intelligibility waning and dialects becoming distinct languages. Inversely, that means that the further you go back in time, the more similar the languages are to one another because, in theory, they all originated from one common ancestor.
The exact opposite has happened in the so-called Altaic languages. The major criticism of the Altaic language family, and the reason why many linguists discredit this theory, is because the languages involved are less similar the further back in time you travel. Even though there are similarities between modern day Turkish and Mongolian and Korean and even Japanese, these similarities diminish or are even non-existent in earlier forms of these languages. This is a huge problem for the theory of the Altaic language family because if these languages are, in fact, related, they should originate from a common ancestor, and their similarities should logically increase, not decrease, as we go further back in time.
Despite this, the debate on this topic is still pretty lively, and I have to admit, there are good arguments on both sides. But what do you think? Do you believe the Altaic languages have merit? Do you speak one of the languages involved? Let me know in the comments below.
Further Reading:
- “Altaic Languages” by George Starostin.
- “A Brief Explanation of the Altaic Hypothesis” by Dr. Cynthia L. Hallen.
The Polyglot Files 
Very good article. Linguistics is a very complex
science and can be very helpful in solving the mysteries
of the past.
However when there is very little harmony with the physical
sciences of archeology and anthropology, theories
become quite controversial. In the case of the Altaic language
family, it is a very weak theory.
While there is physical evidence of Mongolian people in
that area in very ancient times it was quite minimal probably
more novel than anything.
The ancient Jomon people of Japan share much more in
common with a number of American Neolithic Indian tribes and
Tibetans than with the Turks.
I agree 100% that similarities in language and other cultural
aspects should be stronger as you trace back…and in this
case the theories become weaker and more washed out.
LikeLike
The biggest mistake of this article is the section “Why is This Theory Controversial?”
You say ”Besides, just because two languages have cognates or shared grammatical features, it doesn’t mean they are necessarily related”
🙂 Do you really think these are just two languages and have some similarities, thats all?
Their grammer is the same, these people fought each other or against a common enemy, married each other, lived in the same steppes, drank water from the same river, ate the same food, rode the same horses, even their way of singing was the same (talking about the period of 500 and 1200 names)
And you give the example of the English and the Zulu to prove it.
You say ”To prove this, we can look at English and Zulu. Both of these languages have a subject-verb-object word order, and English has borrowed a number of Zulu words such as “impala” and “vuvuzela”.
And so you mean it’s proven?
At this point, I want to ask you a question.
What would you think if an entire zulu village sang (a folk song) with the same accent as a whole village in England?
Or if you had a DNA test and found that you were 20-30% Zulu?
I think your proof has collapsed
The problem here is that the English and the Zulu did not live together for thousands of years, but Turkic and Mongolic.
Yes, Turkic and Mongolic are not the same language, but we cannot say that they are not relatives.
Yes, they borrowed words from each other.
But obviously there was more than borrowing words.
These were related peoples, they lived with each other and they even killed each other.
According to linguistics, if you want to say that these two languages are not related to each other, you have to find the ancestor of both languages, you cannot prove it with zulu and english. Please dont such a simple things again. Dont even think about it.
LikeLike
Your statement about “similar grammar” is not true. Even altaicism of the first wave, which claimed only about the relationship of Turkic, Mongolian and Tungus-Manchu, was smashed to the head because, upon closer examination, the grammar and, in particular, the verbal morphology of Turkic and Mongolian are very different. That is why altaicists of the second wave, such as Starostin, focused on the search for supposedly “common” vocabulary. In order to ascribe “Altai” kinship to Japanese, Starostin put forward the idea that there was no vowel harmony in the “Proto-Altaic”, although earlier it was precisely the presence of vowel harmony that Altaists presented as an “obvious sign of kinship”. Finally, with regard to Japanese, its grammar is ABSOLUTELY different from Turkic or Mongolic. If, from your point of view, grammar is only word order, then I have very bad news for you. The verbal affixes of Japanese are quite obviously grammaticalized auxiliary particles of a once analytical language. In the same way, case meanings in Japanese are expressed analytically, in contrast to all other languages included in the “Altaic” pseudo-family.
Are you talking about DNA tests? The relatedness of populations is determined mainly by Y-DNA haplogroups. In the Japanese, 44% of the population has haplogroup D, which remains from the autochthonous population of the Jomon era, and about 50% falls on the “Han” and “Austro-Asiatic” subclades of haplogroup O (Moreover, it is the “Austroasian” that dominates). This, like the signs of the past analytical system, clearly proves that Japanese is not “Altai” , and its ancestral home was in Southeast Asia. At the same time, the Mongolian peoples and the Tungus-Manchurian peoples are characterized by different (sic!) subclades of haplogroup C, and for the Turkic peoples – mainly R1a, N, even J, only the Kazakhs have haplogroup C.
According to all the laws of logic and science, the burden of proof lies with the assertor. And it is YOU, the Altaicists, who must prove the alleged “kinship” of the Mongolian and Turkic, or the Japanese-Ryukyu and Turkic … But you still have not succeeded in this. All you have done is show religious fervor and deafness. You hypothesized based on “similarity in grammar”, when that was debunked you started saying, “Okay, the grammar is not that similar, but we’ll find a common vocabulary”, and when that “searching” led to the release of absurd anti-scientific bullshit called EDAL, which was refuted and, on the contrary, with its absurdity proved only the weakness of Altaicism, you, Altaicists, simply began to pretend that your fantasies had not been refuted by anyone, and turned into a sect. The Altaic hypothesis is simply a sectarian religion.
LikeLike