On 9 Oct 2016 16:49, "selpahi" <seladwa@gmx.de> wrote:
>
> On 08.10.2016 02:23, And Rosta wrote:
>>
>> Each argument place is either
>> distributive or collective? Would you not also want an "unspecified as
>> regards distributivity"? And wouldn't this mean that where the xorxesian
>> underspecification of distributivity would have one predicate with, say,
>> three argument places, yours would have 2^3 or 3^3 predicates? This
>> looks so untenable that I conclude I must be misunderstanding you.
>
>
> This would indeed be untenable, but I do not believe that you need every version of every predicate. For example, I believe that a distributive {citka} is enough. Very often, a non-distributive version is either not distinct from the distributive version, or includes some added meaning of "doing it together while possibly some of them only watch" (things like {kansi'u lo ka citka}). There is a lot more to be said here, but I'd rather first hear any additional points from you.
My view is the same as xorxes's (thanks to my long ago having been persuaded by his insights on this point, as on so many others). "The guests have eaten all the sausages" is not (fully) distributive and judgements of whether a given predicate, such as "eat all the sausages", can be fully collective, or fully distributive, or intermediate, are more a matter of pragmatics than of formal semasiology.
So I think the better way to handle this linguistically is to have ways to explicitly encode full collectivity ("plurality X but not necessarily any subplurality of X" = ru'o) and full distributivity ("every single X but not necessarily any plurality of them" = ro('oi) pa), without fretting about their truth-conditions with any given predicate.
--And.
--