Re: vowel harmony

On Mon, 2005-11-21 at 19:57 -0500, Kit La Touche wrote:
> it seems to me likely that lexical items preserve a pattern like
> this, which is called harmony for largely historical reasons, but
> that affixation wouldn't be vowel-harmonic, simply because speakers
> wouldn't have enough evidence that it's a harmony pattern when
> acquiring the language to keep it that way.  harrison is doing
> research on this at the moment - i don't know whether he's published
> anything about it yet or not.  he's a specialist in siberian turkic
> languages, tuvan particularly, so i should ask him about modern
> mongolian.  (he teaches at my school.)
I would be interested to hear anything. I'm just basing that on the
Wikipedia article on the same, which is all I know about Mongolian, and
which is rather thin and potentially of dubious quality. It only says
that "all the vowels of the word" must be the same, and doesn't mention
anything about vowel harmony in the minuscule section on grammar. (It
does, however, observe that "short /o/ is phonetically [8]" (i.e.
close-mid central rounded), but given that all vowels exist in
length-distinguished pairs, and nothing is mentioned about the phonetics
of /o:/, one assumes /o:/ is back, approx. [o:].)

> then again, there's finnish, which has frontness-harmony, with the
> vowels [y], [ø], [æ] as front and [u], [o] and [A] as back (pardon
> the mix of SAMPA and IPA).  the vowels [i] and [e] are neutral.  yet
> how would an infant acquiring the language pick up on this?  there is
> a somewhat weak featural way of describing the difference, i guess.
> hm.  musings.
I imagine in large just because that's just the way it is. People seem
to be pretty flexible when it comes to learning the intricacies of
language. Also, it seems to be more than just a lexical-type process
like noun genders, it's rather a neutralisation of phonemes in
unstressed syllables. As I understand it, Finns don't even hear/produce
the difference between [y] and [u] there e.g. "olymia" is for most
people pronounced ["olumpia], not *["olympia] (I gather it's something
along the lines of "lama" vs "yama" for "llama" in English, except
phonologically motivated). No doubt we can blame the absence of /M 7/ in
stressed syllables for the neutral behavior of /i e/. So if Finns don't
pick up on/use [+/- front] in unstressed syllables, and if there is no
[-back +high -rounded +short] phoneme, then [+high -rounded +short] has
no choice but to equal /i/, whereas [+low -rounded +short], which
remains ambiguous, is given a final interpretation based on the
syllable's stressed vowel.

--
Tristan.

Replies