Are there words which sounds very similar in different languages, and which are proven not to be the result of cultural exchange? For example, we know most words related to technology comes from English, so that's probably why computer is computadora in Spanish and kompyuta in Japanese. But is there proof that words coming from cultures isolated from each other having very similar sounds? (not counting onomatopoeias which are supposed to imitate how a noise sounds) . If so, can you give me some examples?

  • 4
    The classic example is Mbabaram and English 'dog', which have the same meaning and similar pronunciation. – WavesWashSands Apr 15 '17 at 17:30
  • 4
    "For example, we know most words related to technology comes from English" I'd say, IT-related technology, not technology in general. – Constantine Geist Apr 15 '17 at 18:01
  • 1
    It is to be expected that for any two languages there will be a number of chance matches (ie words with similar form and meaning). Here's further discussion: zompist.com/chance.htm – Gaston Ümlaut Apr 16 '17 at 9:53
  • 2
    They're all over the place. It's easy to find half-a-dozen words in any two languages that (a) resemble one another phonologically and (b) resemble one another semantically. Yucatec Maya hól 'hole', for example. Larry Trask has a section on this problem in his Historical Linguistics text; he examines the statistics and shows why it happens. The key term is "resemble" -- there are no systematic sound shifts, and the phonological and semantic "resemblances" are at the whim of the analyst, hence they're all over the place. – jlawler Apr 16 '17 at 15:50
  • 1
    It's bound to happen if you relax your criteria for "resemblance", and they are always very vague and never systematic. There are only so many possible roots in a language and only so many things they can mean, so there are bound to be some hits; these can be multiplied by confirmation bias. – jlawler Apr 16 '17 at 15:53

I think False Cognates : https://en.wikipedia.org/wiki/False_cognate might be what you are looking for.

  • false cognates from different languages, yes, that's what I'm looking for. – Pablo Apr 16 '17 at 15:36
  • Happy to help! :) – WiccanKarnak Apr 16 '17 at 15:38
  • @Pablo and as for an example, there is Farsi behtar and English better. – WiccanKarnak Apr 17 '17 at 7:29

There is a Facebook group Linguistic Coincidences & Curiosities the members of which have been collecting such false cognates for years already. My favourites are the Latin and Malaysian "dua" which means "two" in both languages, there are also "reverse" coincidences like in Polish "tak" means "yes", and in Indonesian it means "no".


I will give you a perfect example from Turkish and Japanese. Turkish iyi and Japanese ii いい are read totally same. And more interestingly both of them have the very same meaning good. If you add Ural-Altaic hyptothesis, one can totally jump to the conclusion those 2 words have something in common.

I don't know story of the いい but I can tell about the iyi. In Istanbul Turkish it is iyi, but in most of the Anatolian accents it is eyi or eyü. It comes from the old Turkic edgü->eḏgü->eygü->eyü->eyi->iyi.

So actually both words have nothing in common except meaning.

  • 2
    What about Spanish vs Arabic articles? Given the long history of Arabian domination of Iberic Peninsula, the fact that Latin has no articles, and the striking similarity (el vs al), it seems to beg the conclusion that el is an import from Arabic. And yet, el comes from a Latin demonstrative. – Luís Henrique Apr 19 '17 at 10:48
  • It is very interesting! Do you know any paper or article about that? I checked wikipedia, but they didn't mention any source. – kabraxis Apr 19 '17 at 13:38

Italian strano and Russian stranno, both meaning "strange" and totally unrelated. The English word "strange" is related to Italian one.


The Nahuatl word teotl meaning god makes me think of the ancient greek theos, which has the same meaning [Theos — From Proto-Hellenic *tʰehós (whence also Mycenaean Greek 𐀳𐀃 (te-o)), a thematicization of Proto-Indo-European *dʰéh₁s, s-suffixed noun derived from *dʰeh₁- (“to do, to put, to place”)].

Or even the Nahuatl verb cochi meaning to sleep reminds me of the french verb coucher, which means to lay down (although they are pronounced a bit differently, since the french "ch" sounds more like a "sh") , and is often used as to say to go to bed (e.g. "Je me couche" -> "I'm going to bed").


There are probably thousands of such words. Off-hand, I can give tr bardak/ru бардак, iv מטוס/fr matos.

It would be nice though to get the full list, in any pair (or more) of languages !


  • Because Russia never had significant contact with Turkic languages … – Anton Sherwood Apr 21 '17 at 19:27
  • 1
    Both historically and linguistically your assumption is wrong. Even the word for language in russian is a Turkic word. – kabraxis Apr 23 '17 at 14:18
  • Russian бардак is Turkic borrowing. – Anixx Aug 29 '18 at 10:52
  • @kabraxis Russian word for languge язык is not a borrowing, it is inherited from PIE and cognate to "language" and "tongue". It comes from PIE dnĝhuea̯s. – Anixx Feb 19 '20 at 6:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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