Nortaneous wrote: ↑07 Jan 2022 20:29
OK, so you have a syllable structure of (b d g m n ŋ l)(P F)(m n j w)V(N j w) [...]
No, I don't.
The syllable in question is (C)(G)V(N)(ʔ) , with my arbitrary symbols. The prestem does not matter here. (Of course, it is possible to make it matter but ...)
Why do voiced plosives only appear in preinitial position?
Because they are short and "syllabic". I think that is more or less "natural". They don't, of cource, contrast with the voiceless ones. Because they are voiced on the surface, I see no good reason for a deeper analyses. It probably has an epenthtic vowel in some phenetic contexts.
Is /d.tn-/ allowed? Does it contrast with /d.tʰn-/?
In the initial system all possible combinations are allowed. That's why I find it that elegant.
Are you entirely sure that's your syllable structure?
Not at all, that's why I'm asking. But solving the problem in Gordion style and changing everything is not my intention.
How is the aspiration contrast realized?
There is no phonemic distinction with voicing, so it does not have to be very strong, but the idea is that it is an (post-)aspiration contrast.
Why not just allow (P F)(j w) and (P F N N̥)(j w)?
You are basically suggesting C+glide clusters only?
That is what I do in most of my langs and I find it very boring.
Was there some kind of *r (or *l) such that *mr- *nr- > mn- n-?
Possibly, but they are not synchonically meaningful, at least what I thought.
You could just add an h slot and prohibit hF (or Fh; maybe there's no good way to determine the order). For example, /d.ʰtna d.ʰna d.tna d.na/ [tə̥t̚n̥ʰa tə̥n̥a dəd̚na dəna], or even [tə̥ˀn̥ʰa ˀt̚n̥̩na dəd̚na dn̩a], where the initial prevents the medial nasal from absorbing the preinitial schwa to produce a syllabic consonant, although I think this would imply a dispreference for stressed open syllables of the sort you see in English and Aslian and you'd probably want final stops.
It's also possible for glottals to condition vowel lowering, so maybe you do have these clusters and /d.ma d.m̥a d.m̥ma/ are [dəma dḁm̥a tḁma]. This would be a fairly deep analysis, though, and would need to be justified somehow by some phonological process. (Or maybe just reference to /mn- nm- clusters, which could exist.)
Those analyses are very abstract and should have reasons.
If the prestem consonants had voiceless realizations, it would be much easier to analyse them having voicing contrast.
They are nicely out-of-the-box. It could actually be interesting to utilize those ideas but I don't know much of the morphology yet.
Thank you for pointing out Aslian. That looks very interesting.
I'm not sure if this analogy will make any sense, but the phonetics-phonology boundary, especially under conditions of high compression and complex root structure, is a little like graphics programming targeting composite displays - there are 'smearing' processes that exploit the structure of the signal to intensify contrasts. Think about how the noncontrastivity given by the structure of the signal can be leveraged to increase articulatory ease and/or contrast salience. (For another example, in some English dialects /æw/ flattens to [æ] in specifically the environments in which /æ/ is raised to [eə].)
Couldn't those dialects just have /æ/ and /eə/?