zompist bboard

Hey all!

I was wondering, do languages which show vowel harmony seem to separate the classes of words (based on vowels) by what they mean? (eg: are colours front-vowel words, and days of the week back-vowel words, or something?). Or is the distribution more random?

Thanks!

The distinction is very unlikely to be meaningful in any way; the backness of the vowels is just another phonological feature, like voicing in consonants.

bricka wrote:The distinction is very unlikely to be meaningful in any way; the backness of the vowels is just another phonological feature, like voicing in consonants.

Yeah, for the most part I'm not sure there's any kind of semantic categorisation attributed to vowel harmony, e.g. colour=front, day names=back, etc. However, one example often cited comes from Manchu in which "ama" means "husband" and "eme" means "wife" and similarly "haha" means "man" and "hehe" means "woman" suggesting front vowels are feminine and back vowels are masculine. However, the example texts provided by wikipedia provide a counterexample in "sargan", also meaning "woman", suggesting the male-female thing only occurs sometimes.

Not as far as I know in natlangs, but if you want to develop that sort of thing in a conlang, use classifiers or something, fuse them to the noun, and develop vowel harmony from whichever end they fuse to. For example, if English developed front-back vowel harmony, the days would all have front vowels (since they all end in /de/~/dI/). And if, say, "color" became a classifier /kV/ and fused to the end of the color names somehow, the colors would all have back vowels once harmony developed.

I can see vowel harmony coming to largely align with a formal gender system, which in turn will largely align with natural gender if it's a IE-ish gender system. Let's say that -i is a final vowel which marks the feminine gender (just like Spanish does with -a). If that -i triggers front vowel harmony, and then is lost, suddenly you have a whole bunch of feminine words with front vowel harmony, but the feminine words which happened not to end in -i won't necessarily have it, and masculine words which happened to end in -i (like, the opposite situation of mano in Spanish, which is feminine but has a masculine -o vowel), will likewise have front vowel harmony but still be masculine.

I made a map of the Finnish names of the cities and municipalities in southern Finland. Red: consists of front vowels (takes endings like -ssä/-llä). Green: consists of back vowels (takes endings like -ssa/-lla). (If the name was a compound, I only looked at the last word.)

The distribution isn't entirely random: names that consist of front vowels are rarely found on the coast and are usually clustered together.

Now just account also for the fact that back vowels are more numerous overall.

While there is no general separation according to backness, words with /ö/ (discounting /öy yö/) are almost universally "affectiv" in their flavor. Sometimes there are even corresponding neutral-register words:
henki "spirit, breath" ~ hönki- "to breathe or pant loudly at something"
letku "hose" ~ lötkö "something slack or squiggly"
lokoile- ~ lököile- "to lie about"
mämmi "malted rye pudding" ~ mömmö "sludge"

Finnish may not be the best example however, since vowel harmony is an ancient feature. I'd expect a language that has recently develop'd vowel harmony to display some backness-changing derivation patterns (as we see in eg. Germanic umlaut). There may be some ancient ones still lingering in Uralic too tho (most conspicuously *jalka "foot" ~ *jälki "footprint)

daryush wrote:I was wondering, do languages which show vowel harmony seem to separate the classes of words (based on vowels) by what they mean? (eg: are colours front-vowel words, and days of the week back-vowel words, or something?). Or is the distribution more random?

Korean, which makes extensive use of vowel harmony for purposes of sound symbolism, actually splits colours into two categories according to traditional vowel categories. Those with historically back vowels ("light isotopes", in Samuel Martin's terminology) typically designate lighter and clearer hues than those with historically front vowels ("heavy isotopes").

You said back twice.