Scribo wrote:Finnish is always brought up in these discussions
Well, I always do, since it's my native language and anyway the only living one I know with phonemically distinct short and long vowels that are qualitatively similar...
Hungarian could be just as good an example, but since I don't know Hungarian I'm not bringing it up
(but there are many similarities with Finnish, I believe: distinctive vowel length both in stressed and unstressed syllables, stress on first syllable etc, and the languages are distantly related).
Latin too has distinctive vowel length both in stressed and unstressed syllables with apparently no difference in quality, so the vowel system must have been actually quite similar to Finnish, more so than Greek.
What I'm arguing is that I believe that actual tape recording of actual speech has more or less destroyed the idea that the relative difference in length between short and long vowels is constant, at least for Finnish. This Wikipedia article
says that there are actually four different vowel lengths in Finnish, dependign whether the sounds are in stressed syllables or not - and I suspect that even this is an oversimplification, since other prosodic factors can come into play. This makes me suspect that the same is true for Latin. But of course, what technical equipment shows and what native speakers perceive are two different things, so it just might be that the ancient Romans felt it that way; only I remain sceptical. (Song might be different, because sung notes have to be of a clearly defined length. But the prosody of song and the prosody of real speech are very different anyway.)