Implications regarding vowel harmony?

If you're new to these arts, this is the place to ask "stupid" questions and get directions!
Post Reply
Nachtuil
greek
greek
Posts: 595
Joined: 21 Jul 2016 00:16

Implications regarding vowel harmony?

Post by Nachtuil »

I am playing around with creating a language with vowel harmony. If it matters it is as follows:
+ -
i ɪ
e a
u ʊ
o ɑ

Anyway, I figure vowel harmony in words can only really occurs in languages with agglutination or other compounding occurring, right? Are there any interesting semantic opportunities that arise when it is in place that I may want to be aware of? Is there nothing remarkable perhaps?
User avatar
gach
MVP
MVP
Posts: 513
Joined: 07 Aug 2013 01:26
Location: displaced from Helsinki

Re: Implications regarding vowel harmony?

Post by gach »

If you want harmonic alternation, it's a good idea to have some morphology so that you can get different vowels on affixes when you attach then on different base words. However, vowel harmony is really nothing more than long distance assimilation between vowels. Apart from that definition there's quite a lot of freedom in how the harmony works and what exactly should harmonise with what.

There's this example of vowel harmony in Northern Qiang where the numeral "one" is affected by the vowel of the following classifier. Provided that the numeral and the classifiers are indeed better analysed as independent words, you have here a case where vowel harmony spreads across word boundaries. Alternatively you can say that in such cases there are harmony words that are sometimes longer than the words defined by other criteria like their syntactic behaviour or the placement of stress. Mismatch between the vowel harmony domains and convenient word boundaries is anyway common. In Finnish regular vowel harmony operates between the stem and affixes but not between the components of compound words. Thus you have alakoulu ("lower primary school grades") with both back vocalic ala and koulu but also yläkoulu ("upper primary school grades") with a front vocalic ylä and a back vocalic koulu.

You can in principle also have constraints for what vowels can co-occur in a word even when you don't have any morphology where this would induce vowel alternations. It's perfectly reasonable to call such constraints vowel harmony as well.
Nachtuil wrote:Are there any interesting semantic opportunities that arise when it is in place that I may want to be aware of? Is there nothing remarkable perhaps?
My guess is on there being no overarching semantic correlations at all. More limited sound symbolism does happen but I'm not sure that any such semantic associations would extend to cover whole vowel harmony classes.
ImageKištaLkal sikSeic
Nachtuil
greek
greek
Posts: 595
Joined: 21 Jul 2016 00:16

Re: Implications regarding vowel harmony?

Post by Nachtuil »

Thank you so much for the detailed response Gach. It is good to know that there are a lot of possible ways to implement VH like that. I hadn't really considered affixes being affected while root words inside compound words might not be. It sounds like there is a lot of play in what can be affected by vowel harmony and what can't!
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Implications regarding vowel harmony?

Post by clawgrip »

Jamaican Patois is said by some to have vowel harmony. Seems to be two forms of harmony, the first being which vowels can occur together in a diphthong, and the second being which vowel the standard English schwa is realized as (since there is no schwa in Jamaican Patois). I am no expert and from what I can tell, there are those who dispute it, but still, whether or not you accept it as an accurate description of Jamaican Patois, it still provides at the very least a theoretical example of how a non-agglutinative language could have vowel harmony.
Nachtuil
greek
greek
Posts: 595
Joined: 21 Jul 2016 00:16

Re: Implications regarding vowel harmony?

Post by Nachtuil »

Oh! That is so interesting, and so close to home in terms of it being a dialect(sort of) of English! Thanks for pointing to that Clawgrip!
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Implications regarding vowel harmony?

Post by clawgrip »

There's also the most obvious solution of just having multisyllabic morphemes that don't mix vowels from different classes, even if affixes are minimal. Another possibility is having expressive words that vary their vowels from one class to another in order to shift the meaning slightly (Korean does this, if I'm not mistaken).
Nachtuil
greek
greek
Posts: 595
Joined: 21 Jul 2016 00:16

Re: Implications regarding vowel harmony?

Post by Nachtuil »

clawgrip wrote:There's also the most obvious solution of just having multisyllabic morphemes that don't mix vowels from different classes, even if affixes are minimal. Another possibility is having expressive words that vary their vowels from one class to another in order to shift the meaning slightly (Korean does this, if I'm not mistaken).
Interesting! Obvious to you perhaps! :P Perhaps vowel harmony could be a way of marking tense or aspect or something in verbs, as in, aspect one takes from vowel set 1, aspect 2 from vowel set 2. That may be contrived, but I imagine it would work so it would be interesting to see if that exists.
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Implications regarding vowel harmony?

Post by clawgrip »

Obvious in the sense that the basic idea of vowel harmony is that vowels of a certain class cannot occur in the same word as vowels from another class. Even if the language is completely isolating, this is still perfectly doable if you have multisyllabic morphemes.

Switching vowel classes as a grammatical marker seems a bit iffy, since it would require all words to be of one vowel class by default so that they can all change to the other class when conjugated. But if all words are of the same vowel class, then it kind of stops being vowel harmony. Switching classes to me really seems more suited to semantic rather than grammatical variation.
Nachtuil
greek
greek
Posts: 595
Joined: 21 Jul 2016 00:16

Re: Implications regarding vowel harmony?

Post by Nachtuil »

clawgrip wrote:Obvious in the sense that the basic idea of vowel harmony is that vowels of a certain class cannot occur in the same word as vowels from another class. Even if the language is completely isolating, this is still perfectly doable if you have multisyllabic morphemes.

Switching vowel classes as a grammatical marker seems a bit iffy, since it would require all words to be of one vowel class by default so that they can all change to the other class when conjugated. But if all words are of the same vowel class, then it kind of stops being vowel harmony. Switching classes to me really seems more suited to semantic rather than grammatical variation.
Ah I think I see what youre saying about vowel harmony not really fitting or making sense for such a use compared to verb classes. Thanks again.
Ælfwine
roman
roman
Posts: 940
Joined: 21 Sep 2015 01:28
Location: New Jersey

Re: Implications regarding vowel harmony?

Post by Ælfwine »

It's worth noting that Old Norwegian had vowel height harmony, and Old Norwegian wasn't agglutinating either.
My Blog

A-posteriori, alternative history nerd
Nachtuil
greek
greek
Posts: 595
Joined: 21 Jul 2016 00:16

Re: Implications regarding vowel harmony?

Post by Nachtuil »

Ælfwine wrote:It's worth noting that Old Norwegian had vowel height harmony, and Old Norwegian wasn't agglutinating either.
Thanks, I will check that out!
shimobaatar
korean
korean
Posts: 10376
Joined: 12 Jul 2013 23:09
Location: UTC-04:00

Re: Implications regarding vowel harmony?

Post by shimobaatar »

gach wrote:In Finnish regular vowel harmony operates between the stem and affixes but not between the components of compound words. Thus you have alakoulu ("lower primary school grades") with both back vocalic ala and koulu but also yläkoulu ("upper primary school grades") with a front vocalic ylä and a back vocalic koulu.
When attached to compound words with both front and back elements, do affixes take their front or back forms? As a side note, are there any affixes in Finnish that aren't suffixes?
User avatar
gach
MVP
MVP
Posts: 513
Joined: 07 Aug 2013 01:26
Location: displaced from Helsinki

Re: Implications regarding vowel harmony?

Post by gach »

Suffixes follow the harmony group established by the last root in the compound, see for example the elative ending -stA in alkuyöstä ("at/from the start of night") and yläkoulusta ("from the upper grades").

There aren't any terribly good examples of prefixes in Finnish. The best contender is probably the nominal negative element epä which appears prefixally very much like un- in English; see aito ("authentic") ~ epäaito ("unauthentic") or pätevä ("qualified") ~ epäpätevä ("unqualified"). However, even these words might be better described as compounds where the first element is an independent derivation of the negative verb e- > epä. Note that there is also an independent verb epäillä ("to doubt") where this element is the sole lexical component, further fitting together with a compound analysis for the epä words. And relevant for your question, there's never harmonisation between epä and the following nominal stems.
ImageKištaLkal sikSeic
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: Implications regarding vowel harmony?

Post by clawgrip »

While we're on the subject, let's not forget regressive vowel harmony, where the suffix rather than the stem determines the vowel class. I seem to remember reading about a certain language with only three vowels, /a i u/, where the vowel of the suffix would be /a/ or /u/, and the harmony would proceed backwards into the word, turning every vowel into the same vowel as the suffix until it reached the neutral vowel /i/ that blocked further regression. I could be remembering slightly incorrectly, but it was something along these lines.
Edit: Looks like I got the idea right but the details slightly wrong:
Vowel harmony in Jingulu affects both nominal (1a-f) and verbal (1g-h) words. Certain affixes containing high vowels (/i/ or /u/) trigger a raising of the final vowel of the root, if it is low (/a/), to /i/. If the preceding root vowel is also low, it also raises, and also the one preceding that, and so forth. An underlying high vowel in the root will prevent any preceding low vowels from raising.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Implications regarding vowel harmony?

Post by Creyeditor »

clawgrip wrote:While we're on the subject, let's not forget regressive vowel harmony, where the suffix rather than the stem determines the vowel class. I seem to remember reading about a certain language with only three vowels, /a i u/, where the vowel of the suffix would be /a/ or /u/, and the harmony would proceed backwards into the word, turning every vowel into the same vowel as the suffix until it reached the neutral vowel /i/ that blocked further regression. I could be remembering slightly incorrectly, but it was something along these lines.
Edit: Looks like I got the idea right but the details slightly wrong:
Vowel harmony in Jingulu affects both nominal (1a-f) and verbal (1g-h) words. Certain affixes containing high vowels (/i/ or /u/) trigger a raising of the final vowel of the root, if it is low (/a/), to /i/. If the preceding root vowel is also low, it also raises, and also the one preceding that, and so forth. An underlying high vowel in the root will prevent any preceding low vowels from raising.
One could in principle also imagine that this kind of assimilation leads (via Umlaut) to a vowel quality change of a whole word to indicate certain grammatical meanings. There would be a zero-allomorph for words that already have such a value.

Proto Language: no harmony
sakat house
sakat-i house-PL
sukut-i mouse-PL
sakat-it house-ACC
Ancient Language: vowel height harmony
sakat house
sikit-i house-PL
sukut-i mouse-PL
sikit-it house-ACC
Modern Language: word final vowel deletion
sakat house
sikit house\PL
sukut mouse/mice
sikit-it house-ACC
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Post Reply