Is there an example where the quantity of a vowel makes a difference in a syllable that is heavy by position?

For a concrete example, this does happen in Finnish (where long vowels are written as double): 'autonsa' ≠ 'autoonsa'. In Latin I can only think of something at a word boundaries: manŭsmanūs, but the second syllable scans long if the next word starts with a consonant. What I am looking for is something like this but word-internally, not in the last syllable unless it comes with a consonant cluster that renders it long independent of the following word.

A pair like pĕsca/pēsca would be perfect — but with real Latin words unlike these of course. If there is no such minimal pair, any suggestions for something that comes close would be welcome. I'm open to treating the diphthongs ae and oe as , for example.

It is possible to know the quantity of a vowel in a syllable long by position, so it is at least in principle possible to have such a minimal pair.

2 Answers 2


The example I feel most certain about is various forms of sum “be” and edo “eat”, in particular the infinitives esse and ēsse and the third-person singular forms est and ēst.

  • Perfect, thanks! I seem to have had forgotten this pair. It's good to have it recorded.
    – Joonas Ilmavirta
    Apr 25, 2022 at 17:54
  • lustrum 'a mud pit, den' ≠ lūstrum 'a purification ceremony' (prob. from luere 'to expiate') are not to be mixed. To quote Festus from Paulus:

    lŭstra significant lacūnās lutōsās, quae sunt in silvīs aprōrum cubīlia [...] cum eijusdem vocābulī prīma syllaba prōdūcitur, significat nunc tempus quīnquennāle, nunc populī lūstrātionem

  • Mārc(ul)us 'Mark' ≠ marc(ul)us 'a (big) hammer' is a funny one.

  • Tusculum (cf. It. Toscana) ≠ tūsculum < tūs 'a bit of incense' - though the vowel length in the former is a bit uncertain because the Greeks spelled it Τοῦσκλον. Still, it ought to be short judging both by its origin (turs-) and its modern outcome.

  • illex < illicere 'alluring; an enticer' ≠ illēx < lēx 'lawless'.

  • villīs < villus 'shaggy hair, fur' ≠ vīllīs < vīlla 'a country-house, farm'.

These two are minimal pairs when any other sound preceeds, in many types of speech even when nothing preceeds:

  • (h)ostia 'sacrificial victim' ≠ ōstia 'doorways'
  • (h)ostium < hostis 'enemy' ≠ ōstium 'a doorway'

Oh, and of course the most commonly encountered one by far:

  • lectus 'a bed' ≠ lēctus < legere 'having been picked out, read'.
  • 2
    It's not like Attic Greek had a short /u/ to use instead of long ου.
    – Cairnarvon
    Apr 26, 2022 at 5:30
  • 2
    @Cairnarvon Indeed, but for this reason the La /u/ was expressed also as ο and mor rarely υ, at least in non-litearary writings. But I guess here we're dealing with straightforward transliteration instead of sound approximation. Also I was confused about the cirmcumflex at first, but of course this is a straight case of the σωτῆρα-rule, only the circumflex is possible on a long penult followed by a short vowel, even if the 'long' in our case is graphic-only. Apr 26, 2022 at 14:31

Your Answer

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

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