The Attic-Ionic dialects of Ancient Greek underwent a sound change whereby, in a sequence of a long vowel followed by a short vowel, the quantities were switched: -V:V- became -VV:-, e.g. -e:o- > -eo:-.

Has a metathesis of this kind occurred -- or does it occur as a synchronic process -- in any other language?

  • The only language I have ever heard of where it happened more than rarely is (Proto-)Greek. But it must be hard to spot or notice without someone pointing it out...great question!
    – Cerberus
    Oct 8, 2013 at 6:09
  • @Cerberus Do you mean you know of sporadic cases of QM in other languages?
    – TKR
    Oct 9, 2013 at 2:36
  • I intentionally left that bit vague, because the answer will disappoint: I seem to remember reading/hearing about it once or twice as happening in other languages, but I have no idea how or where, sorry.
    – Cerberus
    Oct 9, 2013 at 4:17
  • 1
    You may argue that the -VR- > -RV:- metathesis in South Slavic languages is quantitative in nature. This said, though, I should point out that the Slavic languages had eventually lost its quantitative features and substituted them with qualitative ones.
    – Newbie
    Jul 6, 2016 at 6:56

1 Answer 1


I don't know of any language that has this sort of metathesis in vowels other than Hellenic, and since this has gone unanswered for over five years, I suspect there may not be one. But there are a few related phenomena which are worth mentioning.

Compensatory lengthening is a process that happens in several languages (including Ancient Greek): when a consonant is lost for any reason, the vowel before tends to be lengthened. For example, English "night" originally had a short /i/; when the /x/ after it was lost, the vowel lengthened, giving /i:/ (which became /aj/ in the Great Vowel Shift).

More relevantly, the Littera law ("inverse compensatory lengthening") is a sound change that happened in Latin in the early centuries BC. It's named after the word littera "letters", from earlier lītera; in certain environments, the sequence V:C became VC:. This happened with sonorants as well as plosives, such as flāma > flamma "flames".

  • I'm not sure that "flamma" is an example of the "littera law". Although "littera"-type alternations seem to be fairly poorly understood, this summmary by Michael Weiss says on page 7 that "flama" is probably just a "defective" spelling, which I understand to mean one where single M is used to write [mː]. The etymological explanations that I've read indicate that the word would be expected to have originally had a geminate; it's supposed to come from complete assimilation of the G in flăg- (as in flagro) to M. Jan 11, 2019 at 1:38
  • @sumelic Fair! I was going from this paper but your evidence is convincing.
    – Draconis
    Jan 11, 2019 at 1:53

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.