Ancient Greek had two (*) different types of accent on long vowels: the "circumflex" accent indicates high tone on the first mora, and the "acute" accent indicates high tone on the second. (Short vowels only have one mora, so they don't have this distinction.)

Are there any words that are distinguished solely by these two accents? In other words, is there a valid, attested Greek word, that changes into a different valid, attested Greek word, if an acute is swapped out for a circumflex (or vice versa)?

(*) I'm ignoring the grave, since to my understanding it indicates the lack of an accent, not the presence of one. It's also a predictable "allophone" (allotone?) of the acute.


4 Answers 4


Two examples come to mind:

λῦσαι (aorist masculine imperative 2nd person singular, or aorist active infinitive, of λύω) contrasts with λύσαι (aorist active optative 3rd person singular of the same verb). (Smyth's grammar)

And this minimal pair isn't strictly between two different words (since γαλήν᾽ is an apocope of γαλήνα), but it's worth mentioning the oft-quoted mistake by an actor playing Euripides' Orestes (279).

The line goes, ἐκ κυμάτων γὰρ αὖθις αὖ γαλήν᾽ ὁρῶ, "After the storm I see calm waters." The actor mistakenly changed the accent to γαλῆν ὁρῶ, "I see a weasel."


In the comments, Alex B added another example: φώς "man", but φῶς "light" (that is, φάος with Attic contraction). As he put it, "this shows btw the importance of morae for stress placement".


I have always thought there were many such minimal pairs, e.g. ἦ "verily" v. ἤ "or".


I came across another example today: μῆτις, ιος vs μή-τῐς, μή-τι. It may not strictly qualify, as there also is a difference in vowel quantity for the final syllable, but in any case, As both their final syllables are short, it should qualify, though as pointed out, one of them is listed as a compound word. In any case, here are their definitions:

μῆτις, ιος, att. ιδος, ἡ, dat. ep. μήτῑ f. μήτιϊ, acc. μῆτιν, 1) Klogskab, Forstand, Hom., Pind. 2) klogt Raad, Anslag, Hom. o. a. D.      [Tethys, Hes. o. F.
1) wisdom, sensibility […] 2) wise council, estimate, Homer and other poets      [Tethys, Hes. and the following

Μῆτις, ιδος, ἡ Datter af Okeanos o.
Daughter of Okeanos and

μή-τῐς, μή-τι, gen. μήτινος (s. μή), at ikke Nogen, ingen, lgs. μή, μήτις μοι νεμεσήσεται, bare ikke Nogen skal dadle mig; ὃπως με μήτις ὂψεται βροτών, Hom. o. A. (skrives ogs. adskilt) […]
that not anyone, no-one, similarly μή, μήτις μοι νεμεσήσεται, that not anyone shall blame/discipline me; ὃπως με μήτις ὂψεται βροτών, Hom. and others (also written separated) […]

―Berg, C. rektor: Græsk–dansk Ordbog til Skolebrug, second shortened edition, second print, Gyldendalske Boghandel – Nordisk Forlag, København, 1950, p. 484.

  • 2
    The vowel in the final syllable is short in both words, so it certainly qualifies in that sense. (Though since μή τις is transparently composed of two words and the ancient Greeks didn't use word spacing, it's hard to say if they would have considered it one word or two.)
    – TKR
    Dec 9, 2021 at 17:27
  • Oh, it is‽ I assumed that when it was marked as short in one of the words and not the other, it was long. Are there any conventions I need to be aware of here that are at play?
    – Canned Man
    Dec 16, 2021 at 17:41
  • 1
    There's a rule of Greek accentuation that says that when the vowel of a penult is long and accented, its accent depends on the length of the vowel in the ultima: if that's short, the penult is circumflexed; if it's long, the penult has an acute. So the circumflex in μῆτις suffices to show that the ι is short. μή-τις is an exception to the rule because it's really a sequence of two words, which is why the length of the ι has to be explicitly marked.
    – TKR
    Dec 16, 2021 at 17:56

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.