Thank you! I had absolutely not thought of those differences in meaning being important, and they totally probably are.
The system is... complicated. I've coredumped on the Imzy conlang comm for a few days now, but I don't have one nice language page to point anyone at.
Suffice to say, this is a simplification. It's polysynthetic, want that up front. Stressed syllables are basically glottalized and retracted and indicate the patient of a verb. Because all core arguments are generally marked on the verb: topic, agent, and patient, with topic often conflated with one or the other. Agent is optional in some circumstances; patient never is.
Unstressed syllables are constrained to neutral tone and equal duration (you can talk fast or slow but all unstressed short syllables should be about the same duration), but long vowels are distinguished and produce a low tone and are produced overlong. They actually suppress stress in their own and adjacent syllables. When this happens, patient marking falls back on inflected forms of the person-marking vowels, which can get quite interesting in fixed vowel verbs. There are a few consistent repair strategies to allow an extra vowel for mutation or stress.
Additionally, there's a vestigial case suffix for patient/accusative that survives in 1st and 2nd person pronouns and some demonstratives that is absent from the 3rd person pronoun and is oddly enough identical to the possessive affix on possessed nouns.
So technically, there's a clearly dead morphological patient-marking strategy, a backup morphological marking strategy, and the preferred prosodic patient-marking strategy. This last, by the way, keeps wanting to migrate to possessed nouns to stress the person-marking vowel in the possessive affix. So far, I haven't decided if I'm going to let it migrate there, but that's my conflicting grammatical systems in a nutshell.
Oh, and also I think of significance is that this particular dialect has lost grammatical number but the 1st person pronouns look very much like a 1st person marker pluralized then nominalized to create a freestanding variation of the 1st person marker. And there are two 1st person markers, but the pronoun looks like the secondary, mostly inherently plural marker instead of the more frequent default 1st person marker. I'd say singular but grammatical number isn't a thing, so it's not actually true. It's just not almost always used for plural like the secondary one is. So it does imply that those pronouns with their vestigial case markers coincided with polypersonal marking on the verb at some point, though the rest of the pronouns have more obscure forms I haven't worked out yet.
And that is not a nutshell, but it exists and it's somewhat concise, so that's good I guess.
From: Alex Fink <[log in to unmask]>
To: [log in to unmask]
Sent: Friday, March 31, 2017 8:39 AM
Subject: Re: Protolanging from an Existing Conlang
On Wed, 29 Mar 2017 21:04:54 +0000, The Scribbler <[log in to unmask]> wrote:
>To simplify my question, is there any good way to tell which grammatical feature is older between two methods for marking the same thing (subtle differences in meaning notwithstanding)? Is it just based on which roots it applies to or is it the more "irregular" seeming it is, the more likely it is to be either older or borrowed? Is there something else obvious I'm missing?
In broad strokes, I can't think of much more you can say. Maybe I'd make minor adjustments, e.g., I can't think of a case where irregular borrowed inflection was retained without pre-existing support for irregular paradigms of the same inflectional category in native words.
Also the subtle differences in meaning shouldn't escape your attention. Grammaticalisation theory says the older method is likely to have the more abstract and general meaning, while the newer method is likelier to retain traces of the concrete meaning its lexical antecedents had.
Also bear in mind that, even if you can say which system is older and which newer, the state of coexistence might itself be very old. Thus English has had both strong and weak verbs ever since pre-Proto-Germanic, when the conversion of the ablauting PIE perfect of root verbs to a preterite le