I speak Japanese, and recently, I've been exposed to Turkish. There's a good deal of overlap between structure, and some words. An example is "good", where it's "iidesu" in Japanese and "iyidir" in Turkish. I looked the languages for the origins and they do not seem to share a common origin, but it sure feels similar. Is this just my brain doing weird sorting or is there actually something in common between the languages?

  • 8
    They are both agglutinative languages, and suggestions have been made in the past that related them to one another – but their shared traits are too superficial and typological for any such suggestions to have held up. In the example you mention, 良い ii [desu] and iyi[dir] may look alike on the surface now, but ii comes from yoi < yoki, whose stem is just yo (-ki being the attributive form of ku-adjectives); while iyi is from eyü < eygü < Proto-Turkic *ed-gü, in which apparently -gü is a derivational suffix. So the stems are yo and *ed, which are not alike at all. May 1, 2023 at 16:45
  • @JanusBahsJacquet thanks for the history. My dialect of Japanese uses よい instead of いい. I had no idea that was the older version.
    – b degnan
    May 1, 2023 at 18:04
  • Japanese and Korean would be in the same family if there were only more cognates -- they're very similar. But if they split, they did so outside the range of the Comparative Method.
    – jlawler
    May 1, 2023 at 19:42
  • 2
    @bdegnan That is indeed an older version. In Old Japanese, it was a ku-adjective, meaning that the final form (終止形) was よし (still used as an exclamation), the continuative/adverbial form (連用形) was よく (still used adverbially), the attributive (連体形) was よき, and the realis form (已然形) was よけれ (still used in the conditional) – but the irrealis (未然形) and imperative (命令形) forms were both the plain stem, よ. Modern-day い-adjectives derived from the attributive form, though the ending /ki/ was weakened to /gi/, then /ɣi/, before finally losing the velar altogether to leave just /i/. May 1, 2023 at 20:31

1 Answer 1


Japanese and Turkish are structurally similar, if you compare them to English, Spanish, Vietnamese or Arabic. They are somewhat more like Arabic, but I am guessing that you don't speak Arabic so you haven't check that connection. Japanese and Turkish are also structurally similar to Quechua, but again that is an obscure language and most people don't go comparing other languages to Quechua, or Vietnamese. A comparison between Navaho and Japanese would reveal strong differences and strong similarities. The explanation for this is that there are only a certain number of ways that languages can be different, so if you look for similarities, you will find them, but if you look for differences, you will also find them. Japanese and Turkish tend to put a lot of propositional content in the formation of single words, using suffixes, whereas English and Vietnamese use distinctive combinations of words to do the same thing. This is basically just a coincidence, that a language has to "decide" whether to have complex words vs. complex sentences.

Another reason for two languages looking similar is that they actually have a common ancestor. We know that Turkish and Kyrghyz have a common ancestor, which explains why roots and affixes in these languages are very similar. In any two languages, you can find some similarity (apart from borrowed words like "baggage"). The similarity that you cite comes in part from picking the most-similar looking word forms of the two languages, where the similarity is coincidental rather than reflecting common historical source. Rather than picking two words from Japanese and Turkish, we would look at the earliest root forms in the two languages, so asking "What is the proto-Japonic root?" (yo) and "What is the proto-Turkic root?" (ed). At that level, the roots are not similar.

To rule out coincidence, one would look for massive and systematic similarity, meaning that we can say what the common original form was and what the rules are that distinguish the English form of the word from the French form (as we have done for the Indo-European, Turkic, Uralic, Semitic, Bantu etc. language families)

  • You could add Tamil and most of the other Dravidian languages to the list. They're all standard SOV languages, with agglutinative (one-dimensional) paradigms and multiple affixes. Plus, the verb comes LAST, and they're all left-branching languages. I was told by a Tamil-speaking linguist who knew Japanese that one could translate morpheme-by-morpheme from Japanese to Tamil, and vice versa. More than half the languages in the world are SOV; it appears to be a very stable saddle point in syntax design. Here's a class handout on agglutination.
    – jlawler
    May 1, 2023 at 19:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.