1. nai pi anu "some of water"
2. náipi anu "some-of water"
3. nai pianu "some of-water"
Right off the bat, I was initially going to say that I feel like it should not be 3, because pianu would literally never appear without a preceding quantity word…but on reflection, I’m not sure why that should necessarily matter. Maybe more telling is the fact that pi can appear without a following noun:
ai sehalu náipi/nai pi?
QU 2S-want some-QUANT
do you want some of?
“do you want some [of this]?”
It is rather convenient to say that -pi is actually a suffix that turns its head into a kind of adverb of quantity. BUT…what if the quantity word is a complex NP? Let’s see…
[[sata lúa-ku tátu]-pi sého]-pi alu ne ka lise...
100 2-x10 3-QUANT bottle-QUANT beer LOC DEF wall
123 of bottle of beer on the wall
"123 bottles of beer on the wall…"
Aside: Wait. Is that how we do partitives?? Bottles of beer, slices of cake? I actually don’t know. Could we have a compound with no particle instead, like Turkish or Macedonian – seho alu? Gotta come back to this!
Anyway, that long number phrase is certainly rather ungainly. But outside of big numbers, I can’t come up with a complex NP as a quantity word that isn’t marginal at best, so it seems that the scope of the pi is just everything, moving backwards, until it hits another phrase with a preposed particle. I think that's okay.
So given that the phrase-final accentuation would be e.g. náipi and not naipí, it does rather seem to make sense to spell things that way. And also, didn't we just announce that -pi turns its host into a kind of pronominal? That might have been the place to start with this whole topic.
An interesting consequence to keep in mind about these theoretical emergent conventions, though, is that the dependent of such pronominal quantifiers would usually be without specifier. This otherwise never happens with NPs in syntax unless they're set off by an immediately preceding particle of some other kind: nekoto "at home," for example. This may have been the reason why pianu felt appealing: that and the fact that it felt like partitive marking in a language like Finnish (lasi vettä "glass of water") which is comfortable, and mirrors the dependent marking which is typical in Koa in other kinds of structures.
I suppose we could say that the -pi suffix kind of turns its host into a giant particle, which we write separately for visual clarity because lásipianu is clearly absurd.
Let's check one thing, though. Imagining a conversation while setting the table...
Ai ípu-pi anu ai lele sa se-si-halu me ka líla-suo?
QU cup-QUANT water QU milk FOC 2SG-ANT-want COM DEF evening-eat
Was it or cupsworth of water or milk that you wanted with the evening meal?
"Did you want a cup of water or milk with dinner?"
Anu, noi-a.
water ask-HON
Water, ask nicely
"Water, please."
I guess that makes sense. But it bugs me that in some other situations, repeating the particle to avoid a bare-stem noun seems to be required:
Ai ne-nomu ai ne-lovo sa se-sano ko se-si-nae ka sími-ni?
QU LOC-upper QU LOC-lower FOC 2SG-say NOM.CL 2SG-ANT-see DEF phone-1SG
Was it in upper or in lower that you said that you saw the phone of mine?
"Did you say you saw my phone upstairs or downstairs?"
...but
*ai ípu pi-anu ai pi-lele
QU cup QUANT-water QU QUANT-milk
Was it a cup of water or of milk
"was it a cup of water or milk"
feels...wrong. Or unnecessary. Or like it's missing the point. Because in the exactly parallel construction of the phone question above, both options actually would be without specifier:
Se-sano...ai Ø-anu ai Ø-lele sa se-halu?
2SG-say QU Ø-water QU Ø-milk FOC 2SG-want
You said...was it water or milk that you want?
"Did you say you wanted water or milk?"
And as I think about it, my example question of "Did you want a cup of water or milk..." was pretty contrived: you really would never say that. You might say "Do you want a glass or a bottle?", but "Do you want a glass or a bottle of beer?" is weird and stilted. These "quantity phrases" (QPs?) really are tight little units, actually quite different from locatives, which is another point in favor of marking them differently.
Also! I almost forgot that the quantified words actually can have specifiers of their own!
Ána-Ø-ni sého-pi ka álu-a-so ce-koa!
give-IMP-1SG bottle-QUANT DEF beer-HON-2PL SUP-good
Give me bottlesworth of the honorable beer of yours most good
"Give me a bottle of your finest ale!"
Compared to e.g. the infamous dependent clauses, this isn't an area I've really considered much through my career in philology, linguistics, or conlinguistics; as such, reaching an coherent understanding of how they would surface in Koa is taking some effort, and I feel clumsy at it. For the moment, though, I think these conclusions are correct, and fairly inescapable once we see that phrases like this are grammatical:
ána-a-ni líma-pi
give-HON-1SG 5-QUANT
kindly give me five of
"please give me five [of them]"
This really is an enclitic, not proclitic, particle.
Back to Turkish and Macedonian: after further reflection, as much as bare nouns may be the default in many languages, including probably the creoles on whose structures we most want to lean, I think we do need to continue to prescribe ípupi anu rather than just ipu anu for "glass of water." Since in N1 N2 structure N2 becomes adjectival and descriptive of N1, the least marked meaning of ipu anu is not "glass of water" but "water glass." This type of ambiguity would cause pandemonium in my household full of literal-thinking neurodivergent people, and is not present in e.g. Macedonian in which we have
чаша вода
glass water
"a glass of water"
but
чаша за вода
glass for water
"a water glass"
On the other hand, returning to the 123 bottles of beer on the wall, eliminating the second pi might sort of make sense given that the focus here is very much on the quantity rather than the nature of the container:
[sata lúa-ku tátu]-pi sého alu ne ka lise...
100 2-x10 3-QUANT bottle beer on the wall
123 of beer bottle on the wall
"123 bottles of beer on the wall…" OR "123 beer bottles on the wall"
The ambiguity feels reasonable in that context. I almost want to ask...though I wince as I do so...whether nested pi-phrases should in fact be forbidden. It may take a little time to answer this, as these kinds of phrases simply don't come up very often in the kind of writing I naturally tend to do. Maybe I need to construct some market haggling dialogs.
Incidentally, the above discussion of accentuation raises the question of whether postverbal locator particles should be written separately after all: ai sehalu mene mé "do you want to come with?", reflecting actual pronunciation, rather than ai sehalu méneme as we currently have the convention. *sigh*.
No comments:
Post a Comment