In Classical times, no distinction was made in writing between /u/ and /w/, or between /i/ and /j/. This distinction seems to have been phonemic, because we see names like Jūlius vs Iūlus.

But are there any true minimal pairs here, where two words with different meanings are distinguished only by a vowel vs a semivowel? If not, what's the closest we can get? (Jūlius/Iūlus is pretty close for i/j, but I don't have a similarly good example for u/v.)


v and u

There are some minimal pairs between [w] and [u]. All of the examples that I have found so far involve words that contain the perfect formative [u] preceded by a sonorant and followed by a vowel:

I am not sure that all of these specific perfect forms are attested.

A small complication is that some words that usually contain [lw] are also attested in poetry with variant forms that contain [lu] in hiatus. I cite a list of examples in this question. When I first read about this, I had the impression that substitution of u for v was a poetic license equivalent (but in the opposite direction) to the variation between syllabic and non-syllabic i or e after a consonant; but now I think it may actually be related to the fact that original -lv- in Latin is supposed to have developed to -ll-, so examples of -lv- in Classical Latin seem to generally derive from processes of syncope that perhaps might not have run to completion in all accents when the relevant poetry was composed.

Some words that usually have [u] in hiatus have variant scansions with [w], such as gen.va for ge.nu.a, but this is not attested with perfect forms as far as I know (related question: Synizesis in perfect tense 'ui').

  • I might be completely mistaken due to my lack of formal training, but isn't qui[kwi]/cui[ku.i] a minimal pair, q/c being the orthographic distinction old Latin made up for this? – Rafael Dec 4 '20 at 10:00
  • 1
    @Rafael Depends if you interpret quī as /kʷiː/ or /kwiː/. Personally I think it makes more sense to analyze /kʷ/ as a single phoneme. – Draconis Dec 4 '20 at 16:32
  • Also, very interesting! Have you done a corpus analysis to find a comprehensive list of these interchanges? – Draconis Dec 4 '20 at 16:34
  • 1
    @Rafael Also evidence within Latin: it only counts as a single consonant metrically (though this one is arguable), other stops can't combine with /w/ in an onset, and it reacts differently with adjacent /u/ (equus pronounced ecus, but servus pronounced servos). – Draconis Dec 4 '20 at 18:52
  • 1
    "the v in solvo is ultimately derived from reduction of the av in lavo" -- I don't think those verbs are related; solvō is generally derived from luō (: Gk. λύω etc.). – TKR Dec 4 '20 at 20:37

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.