Classical Latin stress was famously based on the "penult rule": stress goes on the penult if heavy, the antepenult otherwise.

In later Latin, vowel length seems to have been lost very early: before Late Latin became Romance. But certain sound changes in the Romance languages, like posttonic vowel syncope (deleting vowels in the syllable right after the stress), still rely on the penult stress rules.

So, how long did penult stress last? Did it survive all the way into Romance, even though vowel length didn't—effectively becoming arbitrary? Or did rules like vowel syncope happen earlier, before vowel length had vanished?

  • It is not entirely true that vowel length disappeared; there is plenty of both Late Latin and Mediaeval poetry that demonstrates a naturalised knowledge of Latin vowel length.
    – Canned Man
    Commented Jun 30, 2021 at 14:30

2 Answers 2


I think you're mistaken when you say "certain sound changes in the Romance languages, like posttonic vowel syncope ..., still rely on the penult stress rules".

There are two separate processes involved: the loss of phonemic vowel quantity, and the syncope based on accentuation patterns.

Once vowel quantity ceases to be phonemic, the "penult rule" ceases to operate, but that doesn't change stress patterns, it just means that stress is no longer phonemically determined. Consider, e.g.,

L. virĭdis -> víridis -> Sp. verde
L. catēna -> caténa -> Sp. cadena

The vowel syncope in the first example is a result simply of the stress accent, not on the "penult rule", and can happen whether or not there is still phonemic vowel length. (I believe virdis is occasionally attested in Classical times.)

I believe Christian accent-based poetry is attested from the 3rd century, at any rate, probably indicating the quantitative vowel distinctions were being lost.

  • So if I understand right, you're saying by the third century the vowel length distinctions were vanishing, but the stress fossilized and remained (no longer predictable) until well into the Romance period?
    – Draconis
    Commented Feb 27, 2019 at 16:56
  • @Draconis: yes, although the process of losing quantitative vowel distinction probably took place over several centuries and was likely to have happened at different rates in different parts of the Latin speaking world, and in different social circles.
    – varro
    Commented Feb 27, 2019 at 18:07

I cannot agree with your statement that “vowel length seems to have been lost very early” in Latin. Latin long and short vowels develop differently in the daughter languages. For example Latin short e becomes e in French, but long ē regularly becomes oi (as in habēre > avoir). It is probably true that the distinction was neutralised in the reading aloud of Latin texts, notably in the church, but this illustrates the fact that a reading tradition is not necessarily determined by the spoken language.

  • 5
    With "long" and "short" e, there is supposed to have been a difference in vowel quality which would remain after contrastive vowel length was lost. The reflex of long ē merges with the reflex of short ĭ in French: unless this is analyzed as a lengthening of short ĭ, this merger implies that long ē stopped being a distinctively long vowel. A number of modern Romance languages have qualitative phonemic distinctions like /e/ vs. /ɛ/ without having quantitative phonemic distinctions like /a/ vs. /aː/, /i/ vs. /iː/.
    – Asteroides
    Commented Feb 26, 2019 at 18:42

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.