Why is elision more common than synizesis?

Question

In classical poetry, if two vowels are next to each other (without a consonant in between), there are two ways to avoid the collision:

• Elision removes one of the vowels when the vowels meet at a word boundary.
• Synizesis works within a word, combining two adjacent vowels from different syllables into one syllable.

Neither of these two sound changes is marked in writing.

My experience suggests that when possible, elision happens almost always and synizesis almost never. (I asked a separate question about the frequency of synizesis.) It seems as if hiatus was a bigger problem between words than within a word, and this feels weird to me. I feel that streamlining within a word is more important than gluing words together seamlessly. When studying Latin poetry, I have never seen this issue addressed or even mentioned.

Can you explain why elision is so much more common than synizesis? I understand that "why?" is not an easy question to answer, but I hope that someone has thought and written about this issue although I have not managed to find anything.

Let me elaborate on what I mean by elision being more common than synizesis. (Thanks Joel Derfner for pointing out the need for clarification here.) I do not mean a difference in frequency per verse, but a difference in frequency per possibility. It seems to me that if an elision is possible, there is a 98 % change for it to take place, but the corresponding probability for synizesis is more like 1 %. Don't quote me on the numbers, though, as I have nothing but experience (without statistics) to support them, but they should make the point clear enough. One can also reformulate my question this way: Why, when the possibility arises, does elision happen almost always and synizesis almost never?

If it turns out that synizesis is not as rare as I think, an answer explaining this misconception would be fine. In that case I would appreciate an answer to the linked question about frequency, but an answer without actual statistics is acceptable here.

Answer 1

I haven't read any linguist's comments on this matter, so my post is just a collection of guesses.

Perhaps for grammatical reasons

Elision frequently deletes vowels that are part of grammatical endings, or sometimes part of highly grammaticalized monosyllabic words such as est (with "prodelision"). Grammatical endings often have a certain amount of redundancy, and they also are frequently changed between different forms of a word: e.g. the a at the end of nauta is not present in the inflected form nautīs, and the um at the end of medium is replaced with ō in the inflected form mediōrum. So elision often leaves unchanged the pronunciation of the most recognizable part of a word.

Depending on the context, synizesis might not do that (I'm not sure about that though, so this isn't a complete answer at all.)

Perhaps the combinations of vowels involved are also relevant

The set of vowel sequences that are subject to elision between word boundaries and the set of vowel sequences that can be found in hiatus within a word in Latin are not exactly the same, although there is some overlap.

Many of the most frequent word-final vowels in Latin are not found in hiatus in native Latin vocabulary. At minimum, a, ā, o, ō would fall into this category. (Cser (2016) proposes a synchronic morphological rule of deletion for a and o inside of Latin words when followed by another vowel in derived environments; e.g., in inflected verb forms like first-person singular laudō, which can be analyzed as derived from laudā- plus an inflectional suffix -ō, or inflected noun or adjective forms like neuter plural magna, which can be analyzed as derived from magno- plus an inflectional suffix -a. (p. 96-98))

Depending on whether you treat the endings written with a vowel letter plus M as vowel-nasal sequences or as unitary nasalized vowels, um em im might also count as vowels that don't occur in hiatus inside of Latin words (with a few rare and dubious possible exceptions like circumeo).

Only three vowels are commonly found before other vowels inside of Latin words: e, u, i. These regularly scan short in hiatus (a rule that has certain exceptions), although I'm not sure about the relevance of vowel length and syllable weight to your question. Right now, I can't give a good phonological explanation for why we see hiatus with these vowels inside of words, but not usually across word boundaries.