I've been reading Greville Corbett's new book on Number (2001, Combridge
University Press). I thought I would throw out some of the stuff in the
book to whet some appetites and perhaps start some wheels turning in
people's conlangs. Quenya gets a footnote on page 33 for the -ath plural,
which actually has a natlang counterpart in the South Omotic language Hamer
(spoken in Ethiopia).
There are a rather limited range of number possibilities:
singluar, plural, dual, trial, paucal (a few), greater paucal (a few more),
and general (don't feel like marking number now, even though I could).
There are apparently no quadrals (four) found in the world - he finds
evidence that the claimed quadrals are really paucals.
Rather than simply quantity, some number systems distinguish distributed
all over the place from collected in one place.
There is a heirarchy:
first < second < third < kin < human < animate < inanimate
(personally, I think inanimate should be split to < inanimate < abstract)
If any given point on this heirarchy distinguishes a particular number,
than all points to the left do so too. (standard typology caveats apply,
ie, there are some exceptions)
Now for some of my favorite systems that I've encountered so far.
masc kii tii
fem tii kii
So the number/gender markers swap between singular and plural.
Has a suffix that marks "one more than the logical minimum". Thus, when
attached to second or third person it marks dual, and to first person
exclusive it also marks dual, but to first person inclusive it marks trial.
One more beyond the logical minimum is a normal plural. Ilocano has a
slightly different system.
Both the subject and verb are marked for singular or plural. If both are
marked as singular, you get a singular interpretation. If both are marked
as plural, you get a plural interpretation. If you mark the noun as plural
but the verb as singular, you get a dual interpretation. Similar systems
are found in Zuni and Kawaiisu.
The number morpheme indicates the inverse value, so it changes a singular
to a plural but a plural to a singular. Example, the Kiowa word for 'horse'
is inherently singular, and the inverse marking indicates plurality. On the
other hand, the word for pole is inherently plural, and the same morpheme
indicates singularity in this case.
So, how is number expressed in your conlangs? (No need to repost, Tom.)
Telek only marks number on the verb through subject and object agreement.
Furthermore, only first, second, and animates show number; inanimates do
not distinguish number at all.
fylganal so-pajlo vs. fylganal as-pajlo
The man is big. The men are big.
debaal pajlo debaal pajlo
The basket is big. The baskets are big.
Gassik (a substantial reworking of Igassik, which I bludgeoned the list
with a while back) indicates plurality by reduplicating the final consonant
or vowel-consonant of the noun. This only applies to words that are in the
human category or higher in the heirarchy given above.
nal > nalal 'man(Nom)'
kutu > kutut 'brother(Nom)'
takim > takim 'hand(Nom)'
Unfortunately, or luckily,
no language is tyrannically consistent.
All grammars leak.
-- Edward Sapir