Inverse Grammatical Number

[GoshDiggityDangit]

How does Inverse Number, like the system found in the Tanoan languages, come to exist? Like, etymologically and whatnot.

Spoiler:

If you're unaware, inverse number is a system of granmatical number marking where a noun has a certain "expected" number, eg. single, plural, etc. When a noun is not in this "expected" number, it's marked.
In Jemez, the number-marking suffix is -sh. Nouns fall into 4 number classes. Class 1 is expected singular, Class 2 is expected plural, and Class 3 is expected singular or plural. Class 4 is mass nouns and doesn't get marked.

Code: Select all

               tyó (I) | hhú (II) | pá (III) | aró (IV)
Singular     tyó      hhúsh      pá          —
Dual        tyósh     hhúsh      pásh       —
Plural      tyósh     hhú         pá          —

The table looks terrible but it's 2 in the morning and I've never used the code tables before. Anyway, you get the picture. If a noun is not its expected number, it gets marked.

[Salmoneus]

My guess is that it starts out just by having a plural marker for most plurals, and then a singulative marker for words that are usually plural. My further guess is that coincidentally sound change made those two markers identical. Alternatively, at some point people may simply have recognised that they were marking non-expected number in two different ways and analogised one marker to replace the other.

[Arayaz]

(What Salmoneus said. I'm just here to say I'm almost certain this was the board's 6,666th topic! I like milestone stuff, you know. Also, while I'm not a scholar of Tanoan languages, I do know that every conlang I've seen with inverse number has developed it through a phonological merger of the singulative and plural, just as Salmoneus said.)

[GoshDiggityDangit]

oh that's cool! also yeah it makes sense that it's convergent evolution, i just wasn't totally sure

[eldin raigmore]

Inverse Number is a system of granmatical number marking where a noun has a certain "expected" number, eg. single, plural, etc. When a noun is not in this "expected" number, it's marked.

There could (and IMHO should) be a class of nouns whose expected number is dual. E.g. eyes, ears, hands, parents, ….
If the only number markings are
Expected = 0 (unmarked)
and
Not expected = NE;
That is, if the expected number is singular then dual and plural are marked the same;
if the expected number is plural then singular and dual are marked the same; and
if the expected number is dual then singular and plural are marked the same:

Then how does the language differentiate between the two unexpected numbers?

I can’t tell quickly whether the example you provided is an answer to this question. If it is, I apologize for missing it.

[Visions1]

Maybe it's just that. Inherent expected number is just what they mark.
I love Salm's seconds analysis the best. The first is clever, but unless there's evidence to suggest it, I'm wary.

There could (and IMHO should) be a class of nouns whose expected number is dual. E.g. eyes, ears, hands, parents, ….

I know in Modern Hebrew some words only realy show up in dual. `Enim (eye.pl) and Yadim (hand.pl) are rare, while `Enayim and Yadayim (dual) are the accepted plural. So by marking them as plural, you imply individual eyes or hands, but as dual, you imply a pair, or multiple pairs.
However, in the singular, they're just that.

[VaptuantaDoi]

IIRC some inverse number systems have three types:

Code: Select all

Unmarked:    Marked:
singular     dual or plural
plural       dual or singular
dual         singular or plural

So you might have

mx̌qʷ'elc
eye
"Two eyes"

mx̌qʷ'elc-bim
eye-INV
"One eye" or "Several eyes"

(I can't readily find any natlang examples though)

[sangi39]

A quick read through "Noun class and number in Kiowa-Tanoan: Comparative-historical research and respecting speakers' rights in fieldwork", apparently some of ambiguity in the noun is handled via verb marking, but also set of suppletive verb that use different stems depending on the intended number of the noun