distinction in vowel phonation

[Rosenkohl]

First of all, this is my first post here, so... hello :)

Last week I got overambitious and decided to create an a priori language family. Naturally, I started with the common ancestor. I decided Proto-Conlang would have distinctive breathy voice phonation on vowels, meaning any vowel in any syllable can be either modally or breathy voiced (i.e., /pat/ and /pa̤t/ form a minimal pair).

Is this at all viable? I know Gujarati (a Western Indo-Aryan language) does it, but there breathy voice exists as a result of cheshirisation of a lost /ɦ/*. Other languages seem to distinguish creaky voice, modal voice and breathy voice in vowels, but tone is also a part of the mix. How likely is it that a proto-language would contrast modal and breathy voice as an a priori feature, unassociated with tone?

If it is possible, then what are some of the possible reflexes such a distinction could have in daughter languages?

So far, I've thought of three possible reflexes:

- Cheshirisation on the previous consonant: lax vowel phonation (breathy voice) becomes lax consonant phonation (voiceless, slack voice?), and modal phonation becomes tenser consonant phonation (modal voice, stiff voice?);

[Proto-Conlang /pat/ /pa̤t/ > North Conlang /bat/ /pat/]

- Tone: breathy voice is associated with the high tone, whereas modal voice is associated with the mid tone (in the language that does this, consonants also have an impact on tone);

[Proto-Conlang /pat/ /pa̤t/ > East Conlang /pāt/ /pát/]

- Vowel length (I'm really unsure about this one): modally voiced vowels become long vowels, breathy voiced vowels become short vowels... ?

[Proto-Conlang /pat/ /pa̤t/ > South Conlang /paːt/ /pat/]

Does this sound plausible? What would other interesting possibilities be, attested or otherwise?

Thank you for you help and remember not to bite the newbie (though I am quite delicious).

* I got this information from Wikipedia, which in turn cites Mistry, P.J. (1997), "Gujarati Phonology", in Kaye, A.S, Phonologies of Asia and Africa, Winona Lake: Eisenbrauns. I hope it's accurate.

[QuantumWraith]

Always rememeber that a proto-language is no different than any other language. In the real world, it is never a priori unless it becomes a spoken auxlang. Wherever your proto-language is spoken there will most certainly be a hypothetical pre-proto-language that it developed from. If cheshirisation of /ɦ/ can cause a breathy vs. modal distinction in Gujarati, who's to say the same did not occur in such a pre-proto-language?

As for your sound changes, they seem plausible to me, but I'm no linguist. Though, it seems more likely to me that breathy voice would shift to a lengthened vowel while modal voice would remain unitary rather than vice versa.

Oh, and welcome! :)

[Rosenkohl]

Thanks for you reply :)

I'm aware that there's no special set of rules that apply to proto-languages only. My problem with the cheshirisation path to having a modal-breathy voice distinction was mostly one of distribution. If breathy voiced vowels were all the reflex of a neighbouring /ɦ/, that would limit their occurrence to environments where consonants can appear, which is a concern in the case of Proto-Conlang, with a CLVC (L=trill, lateral, glide) syllable structure. That's why I asked about the likelihood of its being an "a priori feature", that is, one that, as far as one can tell, did not evolve from any particularly intricate set of sound changes.

(In more concrete terms: I want both, say, /pras/ and /pra̤s/ to be possible; with breathy voice arising from lost /ɦ/, because of phonological constraints that I'm not willing to let go of, it would only be distinctive in ɦV(C)- and (C)Vɦ-type syllables in Pre-Proto-Conlang, which, in Proto-Conlang, would give, respectively, V(C) and (C)V syllables... it's not a bad thing per se... but that's why I'm asking for alternatives anyway).

About the vowel length thing, I suppose you are right.

[kanejam]

I usually don't worry about plausible sound changes and constraints for conlangs with no 'attested' ancestor i.e. my a priori conlangs. It is just left as an exercise to the reader to see whether it the changes may be plausible.

With your pre-proto-conlang, you have quite a few options. You can change up the phonotactics of your proto-lang, although you said you are unwilling to do so. You could increase the phoneme inventory of the pre-proto-lang to make more phonemes that might have caused breathy voice to arise; glottal stops can do that too I believe, or maybe even aspirated consonants? Lastly, and this is my favourite, just have your pre-proto-conlang distinguish breathy-voiced vowels as well. There's no rule on how old such a phonation distinction might be; it could have arisen 2000 years before you proto-lang, when the language looked quite different.

Remember that you don't have to be able to explain everything about your conlang and even pre-pre-pre-proto-langs will have irregularities.

[Curlyjimsam]

Is there any reason not to keep the Proto-Conlang constraints whilst changing those of Pre-Proto-Conlang? That is, would it be a bad thing to say that Pre-Proto-Conlang had a more complex syllable structure ((C)(C)V(C)(C) or something), but by the time of Proto-Conlang all consonant clusters except initial consonant+trill/lateral/glide were reduced to a single consonant - the only relic of this being breathy voice where the lost consonant was /ɦ/?

[kanejam]

There is certainly no reason why the pre-proto-lang has to have the same phonotactics as its descendent.

[Rosenkohl]

Certainly not, my concern was mostly about distribution of breathy voice... for it to be as frequent as I'd like it to be for the daughter languages to have the distinct feel I'd want each of them to have, /ɦ/ would need to be much more frequent than other consonants (say, if I want 40pc of vowels to be breathy voiced, that'd mean a whole lot of /ɦ/ in Pre-Proto-Conlang). Or I'd need a different path to breathy voice altogether. The question still is "is there a way for a language to have a distinction in vowel phonation that isn't just the reflex of one lost phoneme?"

Thank you all for your replies :)

[sangi39]

This was an interesting find on the Wikipedia entry for breathy voice:

In Portuguese, vowels after the stressed syllable can be pronounced with breathy voice.

(source: Callou, Dinah. Leite, Yonne. "Iniciação à Fonética e à Fonologia". Jorge Zahar Editor 2001, p. 20)

That means you could have something like:

'a.pa vs a'pa >
'a.pa̤ vs. a'pa >
a'pa̤ vs. a'pa >
'pa̤ vs. 'pa

In this system, stress, originally, is a phonemic feature, with post-tonic vowels pronounced with breathy voice. Later on, stress becomes universally word-final and initial unstressed syllables fall out of use (meaning breathy voice can occur in initial syllables).

That's just a basic sketch but hopefully it provides you with an extra option

EDIT: If anyone knows if breathy voice can derive from creaky voice, then you can derive this from glottalisation. Therefore, if you derive breathy voice in a straight forward manner from older /ɦ/, you can derive glottalised vowels from older /ʔ/.

This means, initially you're pre-proto-language could have had glottalised vowels and breathy voice, with glottalisation shifting to creaky voice, later merging into breathy voice in the proto-language. In the pre-pre-proto-language you could have had /ɦ/ (source of breathy voice) and /ʔ/ (source of glottalised vowels).

You could even derive the /ʔ/ from other plosives, e.g. /kt/ > /ʔt/ between vowels, etc., meaning you wouldn't have to rely on an abundance of /ʔ/ in the pre-pre-proto-language

[Rosenkohl]

That's a very interesting possibility, thank you!

And by the way,

You could even derive the /ʔ/ from other plosives, e.g. /kt/ > /ʔt/ between vowels, etc., meaning you wouldn't have to rely on an abundance of /ʔ/ in the pre-pre-proto-language

I originally did this for the sound changes from Proto-Conlang to one of the daughter languages. There, /ʔ/ derived from glottal reinforcement of long aspirated consonants (/tʰː/ > /ʔt/) and plosive-plosive clusters (/kt/ > /ʔt/). In a later stage, I had these pre-glottalised stops become ejectives (/ʔt/ > /tʼ/). Does that sound okay?

[sangi39]

Seems fine to me