Research CLDR's singular/plural/... subtleties and decide what to support

CLDR has a framework for describing how words vary depending on multiplicity: where English merely distinguishes singular (exactly one) from plural (any number but one), other languages draw other distinctions (e.g. few vs many; or the "one" form may apply to "one plus any multiple of ten", for example) in their grammatical bendings of words.
We'll need a proper understanding of how this works before we can think about adding and using API that's aware of it and the CLDR data needed in support of this.