As an English-speaking student of Yiddish, I was pleasantly surprised to discover that the indefinite article was an before vowels and a before consonants, just like in English. But as far as I can tell, Yiddish developed this independently from English. That is, both turned the Proto-German *ainaz ("one") into two separate words, the number one (eyn in Yiddish) and an indefinite article which is identical in the two languages.

Are there any other examples of two languages which shared a root, separated, and ended up with nearly identical words much later, despite otherwise diverging considerably?

  • 2
    Well, there is a famous example of an indigenous australian language in which the word for dog is 'dog', pronounced the same way as in at least some English dialects: en.wikipedia.org/wiki/Mbabaram_language#Word_for_.22dog.22 My guess would be that placing a consonant there would be preferred based on phonetics. Saying 'a orange' is slightly annoying, since you have to put something like a glottal stop in between unless you want to run the two vowels together.
    – Alan H.
    Commented Sep 14, 2011 at 4:14
  • 7
    Actually, the consonant went the other way around--i.e., it used to be both "an cat" and "an apple," and the "n" was lost before consonants. But my question is about two languages experiencing parallel changes, rather than randomly having similar words. Commented Sep 14, 2011 at 4:40
  • 1
    @Alan: Anschel is specifically asking about languages with a shared root, which rules out Mbabaram, otherwise one of my favourites too of course. Commented Sep 14, 2011 at 9:39
  • 2
    should we discourage this kind of "examples" questions?
    – Louis Rhys
    Commented Sep 14, 2011 at 12:32
  • 2
    @AlanH.: Mbabaram dog and English dog don't have a shared root. The Mbabaram word is related to words in other Australian languages like Dyirbal guda and Yidiny gudaga. The English word comes from an Old English word meaning "muscle". If the two languages have a common ancestor it must be in Africa at least 70,000 years ago, about 20 times further back than we have evidence of any language at all. Commented Sep 26, 2011 at 10:27

5 Answers 5


My favorite example is the convergent evolution of do support. In Modern English, auxiliary do must be used in questions and negative sentences:

(1) John did not kiss Mary
(2) Who did John kiss?

Beninca and Poletto (1998) find that a northern Italian Romance dialect has developed a system of do support very similar to the English one (though it applies only to questions). This pattern is not attested in any world languages other than these two.1 I'm unclear on the etymology. The OED says English do comes from the "Aryan [=PIE? –AE] verb stem dhē-, dhō-" which it relates to the second element in the Latin verbs ab-dĕre to put away, con-dĕre to put together, dē-dĕre to lay down, but not to facere, the source of the Italian word for do. On the other hand, Wiktionary traces Engl. do and Lat. facere to a common root (PIE *dheh1-), but without listing any sources.

1: If you know of another example, me and my research would love to hear about it!

  • 2
    I guess this is what I was looking for--in retrospect it wasn't a very good question, but this is a very good answer. Commented Sep 29, 2011 at 2:32
  • 2
    Middle Welsh has a construction rather like do-support, though not just in negative and interrogative contexts. Some argue that English acquired the pattern from a Celtic substrate.
    – Colin Fine
    Commented May 1, 2012 at 23:45
  • There's no doubt that do and facere are cognates (from PIE *dheh1-); initial *dh- > f- is a regular change in Latin. (The source of the -c- is something of a problem, but that's a different story.)
    – TKR
    Commented Aug 14, 2014 at 16:45

From Germanic languages, both High German and English developed the same diphthong au in the words Haus/house and Maus/mouse independently, while Low German keeps the original long u.

English and High German also both diphthongised the long i to ai, e.g., in mein/my where Low German keeps the original long i.


Some examples taken from the Wiktionary appendix on the Proto-Indo-European roots:

  • Latin: pater; Greek: patēr; PIE root: *ph₂tḗr (father)
  • Russian: пизда (pizdá); Lithuanian: pyzdà; Latvian: pīzda; Polish: pizda; PIE root: *pisd-eh₂- (vulva)
  • Umbrian: pumpe; Welsh: pump; PIE root: *pénkʷe (five)
  • Latvian: spenis; Old Prussian: spenis; PIE root: *psten- (breast)
  • Greek: τρέμω (trémō); Latin: tremō; PIE root: *ter- (to shake, tremble)
  • Illyrian: tri-; Sanskrit; Welsh: tri; Sanskrit: tri; Hittite: θri; Albanian: tre/tri. PIE root: *tréyes, tri- [prefix] (three)
  • Old Irish: dia; Portuguese: dia; PIE root: *déi-no- (day)

And the list goes on and on.

  • 3
    Aren't these examples of shared ancestry rather than convergent evolution? Commented Sep 27, 2011 at 13:22
  • 2
    That's what the OP wants: "examples of two languages which shared a root, separated, and ended up with nearly identical words much later" Commented Sep 27, 2011 at 13:32
  • 2
    Of course now I have to vote up your answer! (-: Commented Sep 27, 2011 at 17:02
  • 4
    Most of these example did not significantly "separate" and then converge: they remained pretty much the same throughout their history. Even the Umbrian and Welsh case, where they independently developed from "kwenkwe" does not show this diverging and then reconvergence.
    – Colin Fine
    Commented May 1, 2012 at 23:50
  • 1
    In many of these examples, the common ancestor was not PIE, but a more recent proto-language (Proto-Balto-Slavic, Proto-Baltic etc), hence some of them are not convergences, but inherited similarities.
    – alephreish
    Commented Mar 2, 2017 at 14:00

I saw a video a while back by YouTuber Artifexian (https://m.youtube.com/watch?v=PBz-JT00MZs) which presents evidence that many languages seem to evolve some similar features based on geography. The most detailed example he mentions in the video is that ejective consonants, where the sound is produced only by the throat and mouth without extra air from the lungs, tend to appear in languages native to high-altitude regions where it might be helpful not to lose excess moisture by breathing into the consonant sounds. Although he doesn't mention this directly in the video, two of the areas where ejectives are common are in the Andes range and the Rockies, and it is commonly believed that native American languages all evolved from the languages spoken by the first Siberians to cross the Bering land bridge. So although the languages of the Andes are distant "cousins" to those of the Rockies, there is a sizable gap between the areas and they probably developed ejectives independently.


Enlish strange and Russian stranno, with the same meaning; English stranger and Russian strannik, with the same meaning. But English and Russian words derived from different and unrelated roots.

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.