greek vowels

[Cognomen]

Why does Ancient Greek use different glyphs to distinguish length for some vowels (e and o) but not others (a, i, and u)?

[Soap]

I believe it was originally a seven vowel setup, with the ē and ō vowels tracing back to originally distinct vowel qualities, whereas the ā ī ū are less common and arose from secondary processes of lengthening. e.g. /oso/ > /oo/ > /ou/ > /ū/ (but still spelled "ou").

I believe that other languages with vowel length often also have separate origins for each of the long vowels. e.g. Japanese has ō far more common than ā, and ō comes from two vowels that were different in quality (one similar to pre-Greek omega, the other to omicron), but there was never a distinct "ō" row in the kanji, let alone one row for each of the two distinct ō's, showing that sometimes the devisers of an orthography are attracted to symmetry.

[Sumelic]

I believe it was originally a seven vowel setup, with the ē and ō vowels tracing back to originally distinct vowel qualities, whereas the ā ī ū are less common and arose from secondary processes of lengthening. e.g. /oso/ > /oo/ > /ou/ > /ū/ (but still spelled "ou").

I believe that other languages with vowel length often also have separate origins for each of the long vowels. e.g. Japanese has ō far more common than ā, and ō comes from two vowels that were different in quality (one similar to pre-Greek omega, the other to omicron), but there was never a distinct "ō" row in the kanji, let alone one row for each of the two distinct ō's, showing that sometimes the devisers of an orthography are attracted to symmetry.

I don't think Greek started from a seven-vowel system like /a e ē i o ō u/. In Attic Greek eta came from fronting/raising of Proto-Greek ā, but apparently in some environments this became ā instead. I think omega just came from original long ō, mostly. The fact that length of a, i, u/y was not written might have to do with the relative frequencies of different long vowels, as you mentioned. More on the historical development of the Greek vowel system here: http://ling.umd.edu/~bridget/pdfs/attic.pdf

My understanding is that Japanese long vowels come mainly from diphthongs, with another source being coalescence of vowels that were originally separated by f/h/w. The way they're written reflects this origin. For example, ō was originally written with two kana as "ou" or "au" (in most cases) because that was the historic source of the vowel. After orthography reforms, "au" is no longer used for this purpose.

[M Mira]

Is it known when the length distinction and diphthongs arose? I'm thinking maybe that the diphthongs monophthongized first in some dialect, then when other dialects developed long-omicron and long-epsilon phonemes, they picked up the conveniently available omicron-iota and epsilon-iota digraphs, but nothing similar exists for long-alpha, long-upsilon, or long-iota, so they made do with the old ones.

If you mean the epsilon-eta and omicron-omega pair instead, then the answer is simple: they're different in quality.

[Frislander]

The Wikipedia page gives lots of information, and it's also worth checking out the other dialects and the Koine Greek page to see the development.

[Sumelic]

If you mean the epsilon-eta and omicron-omega pair instead, then the answer is simple: they're different in quality.

Oh, that is a good point.