dunomapuka wrote:Conlanging on the fly...
Thanks for the input! It looks like a delicious mess that wants a good portion of analogy, to be applied differently in each dialect. Only... it turns out I'm not all that happy with your scenario, because:
- My original idea for O Ayōndui noun phrase syntax is by far the easiest to accomplish if all the suffixes are phrase-level clitics in PL, and not at all compatible with the agreement version as far as I can see. It's typologically much more plausible to grammaticalise a clitic into a true suffix with or without agreement than it is to degrammaticalise a suffix into a clitic, especially when there's agreement involved.
- The locatives as clitics seem to work well (except that the phrase "near the sleeping baby" probably won't ever appear in the accusative except when there's an additional
-ab involved - "the one near the sleeping baby" -, in which situation agreement on "sleeping" and "baby" feels quite clumsy to me, but that's an aside). I also like the dummy pronoun; it might grammaticalise as an alternative to the demonstrative
-ŋu in some way. However, having the
innermost suffixes behave like clitics when the outer ones are true affixes is quite weird. Usually it's the other way around. If the locatives behave as clitics, I would expect all markers following them in the template to only ever appear following them in the phrase.
- The optional
-ŋu-doubling makes sense in the context of an agreement scenario, but not so much with the locatives being clitics, and also not so much with the quantifiers appearing either on the noun or on the dummy but not doubled.
- Disregarding the locatives and the accusative for a while, your scenario theoretically allows the following orders of noun, adjective/gerund, quantifier and demonstrative, if all are present:
Adj N-Qu-Dem
N-Qu-Dem Adj*
N-Qu-Dem Adj-Dem*
N-Dem Adj Qu*
N-Dem Adj-Dem Qu*
However, apparently there's
a linguistic universal that states that post-nominal modifiers can only ever appear in the orders N Dem Qu Adj or else N Adj Qu Dem, so all of your options with post-nominal adjectives (those with a star *) violate that universal. (Also note that N Adj Qu Dem, which is typologically common, is the exact order that would result if all Proto-Lukpanic noun suffixes were phrase-level clitics.)
((This exact universal is already broken in Akana, albeit in a different manner, by Fáralo's Qu Dem-N Adj order.))
So my suggestion would be to rethink that scenario a bit. Maybe a good alternative (which is more in line with what I know about possible grammaticalisation paths) could be something like the following:
1. All suffixes normally act as clitics. When the noun itself is the last word of the noun phrase, they attach to it directly:
naəpaə=mea=ŋu=li "all these babies (ACC)". When there's a post-nominal gerund or locative adjective, they instead attach to a dummy pronoun:
naəpal kpaənui su=mea=ŋu=li "all these sleeping babies (ACC)", lit. "the sleeping babies, all these ones"
2. When there's only one non-locative modifier, the dummy pronoun is not necessary:
naəpal kpaənui={mea/ŋu/li}
2a. In some but not all Lukpanic dialects, dropping the dummy pronoun becomes permissible even with locatives, even with more than one enclitic, or both. (O Ayōndui would be part of this group.)
2b. A dummy pronoun would always be necessary one syntactic level further out when there's a locative clitic that refers to a complex noun phrase as a whole:
naəpal kpaənui (su)=mea=ŋu su=tui "near all these sleeping babies"
3. The demonstrative may optionally be copied to the head noun for emphasis:
naəpaə-ŋu kpaənui su=mea=ŋu=li "all these sleeping
babies"
3a. In some but not all Lukpanic dialects, the demonstrative may also be copied to a modifier word to place emphasis there, beginning to grammaticalise into an inflectional focus marker:
naəpal kpaənui-ŋu su=mea=ŋu=li "all these
sleeping babies". (This may or may not happen in O Ayōndui, not sure yet.)
3b. In some but not all Lukpanic dialects, clitic doubling also becomes permissible with the accusative marker and possibly with quantifiers too, and may later evolve into true case/number agreement. (O Ayōndui would not be part of this group.)
What do you think?