Re: vowel harmony

On Nov 22, 2005, at 5:18 AM, Tristan Mc Leay wrote:
> Yes, I'm afraid to say I'm not sure I completely understood you.
> Perhaps: if a series of changes (borrowings or phonetic drift or
> whatever) in a language cause a large enough proportion of words to no
> longer obey some sort of structured vowel harmony, that language would
> pretty quickly lose its harmony in affixes? So that perhaps if in
> "Starte", a language like Finnish, u: > y: (with y: remaining as is),
> and A: > O: and &: > A:, perhaps, and that brought Starte above the
> cut-off, then Starte would very likely loose its vowel harmony?
>
> If so, that's certainly interesting!
>
> And you're saying this to contradict my conjecture "I imagine in large
> just because that's just the way it is", implying that a language
> could
> have vowel harmony between [i 2: o: 3 a:] in one set and [i: e: u o
> A:]
> in another with [2 e] being neutral. Which probably is a bit
> outlandish
> actually and certainly not something I would wish to imply whilst
> thinking about it! :)
>
> Do I understand? Have I missed something you said, or miss-interpreted
> something?
no, i think you understood correctly.  despite my inarticulateness.

> (PS: Is Uzbek presently losing its vowel harmony, or is it something
> that has happened while Uzbek has been a written language or
> something?)
uzbek (i say this only from papers, no personal experience) has
largely lost harmony - there are still many harmonic roots, but not
enough for the pattern to generalize to affixes; for example, -lar is
the only form of the plural, with no -lar / -ler alternation (if
memory serves).

> --
> Tristan.
kit