[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >
On Sun, Aug 26, 2012 at 10:14 PM, And Rosta <and.rosta@hidden.email> wrote:
> Jorge Llambías, On 27/08/2012 01:40:
> > On Sun, Aug 26, 2012 at 9:01 PM, And Rosta<and.rosta@hidden.email> wrote:
>
> >> 2. I'd like to see a ternary quantifier with a fraction argument (where
> >> the fraction is "all", "some", "two in every three", "a lot" and so
> >> forth).
> >
> > Can't "je" handle this? Let's say "lt" means "a lot", then "la je lta
> > mlta xkra", "a lot of cats are black".
>
> I meant "proportionally a lot; a large proportion of".
What if "lt" means "is a large proportion of"? Then we can say:
la xkra le mlte ltake
Black ones of cats are a large proportion.
Or:
la mlta le lteka xkre
Cats, a large proportion of them are black.
Or:
la le mlte ltake xkra
> >> 3. I would allow variables to be implicitly bound, so that "brda" with
> >> unbound "a" is short for "za brda", where "z" is unary existential
> >> quantifier. This would mean that in "(sa xrma (lu bjrafu vska'aku))
> >> (xkra)",
> >> "xkra" is short for "za xkra".
> >
> > We already allow implicit binding. The proposed rule was that the free
> > variable is implicitly bound by "l" with the same restriction that
> > that same variable was bound last time it was used. Your example is
> > thus equivalent to "(sa xrma (lu bjrafu vska'aku)) ([la xrma] (xkra))"
>
> I'd argue against this interpretation rule for two reasons. First, it
> means that to circumvent it you either have to use overt binding or have to
> search for a variable that has not previously been bound (within what span
> of text?). Second, the speaker-hearer should have to keep track only of
> syntactic binding, not keep a mental note ofall or some
> variable--restriction pairings in the prior text.
What scope do you propose for an implicit s/z binding? If it's minimal
scope, then a single reserved variable should suffice, because each
use will get its own binding except perhaps when used in the very same
predicate, but then we would need a couple more of them. If it's not
minimal, it will be a headache (like Lojban's unbound da).
> OK, I get that if you're going to have restricted quantifiers then {l}
> makes sense as a default restrictive quantifier. And restricted quantifiers
> are certainly needed for stuff like "most", "2 in every 3" and so forth, as
> in (2) above; and restricted quantification for "all" is convenient even if
> not necessary. But instead of "la grka xkra". why not "za je grka xkra"?
> (where z is unary s).
I think it's a prctical way to distinguish what you're talking about
from what you're saying about it.
mu'o mi'e xorxes