Re: "Only" (again) - comments needed so I don't make a fool of myself

[jimc@math.ucla.edu]

>      le  da'apamei        po'u     na'ebo     la mardj citka   lo jintitnanba
>      The all-but-one-some which is other-than Marge    eat/ate doughnut(s).

Very interesting discussion, and I'm happy to see "only" getting
reviewed, and also the suggestion that "just" should be dealt with too
by an authority.  For me, those two words were very slippery.

>From the gismu list dated 12/07/93, N-mei means, X1 is a mass with
underlying set X2 and members X3.  I can easily accept

	le te da'apamei po'u na'ebo la mardj...

i.e. members in extension with Marge restricted out.  When the predicate
is eating donuts, the eaters do it in extension.  Clearly this is wrong:

	le se da'apamei po'u na'ebo la mardj...

i.e. the set which is not Marge.  I think 

	le (-) da'apamei po'u na'ebo la mardj...

is closer to the set version.  Certainly the eaters are still doing it
individually, but { (X)  po'u na'ebo la mardj } can't mean anything
other than "(X) restricted to not be Marge".  Maybe it's an ant colony
which has latched onto Marge's donut.  The point is, (X) and Marge are
in the same semantic category, and are potentially identical.  Think of
the underlying sets; clearly it's bogus (Russell's paradox) to restrict
a set so as to take a subset which is not equal to a designated member.

A possible reply is, the "value" of a mass is its members in extension
(in which case the restriction is OK), with an implied restriction that
the members form a set-with-structure, which is the mass.  This
solution certainly takes care of the semantic conflict, that whatever
the mass (of donut eaters) is doing, the donuts are being eaten by
members in extension. 

But to me, this way out seems like cheating, in that you ought to be
able to talk about the mass as a unit, separate from the members and
from the underlying set, and if the mass necessarily slithers off into
crumbs of members, you can't talk about it at all.  In any case, you
can expect some comments about massification from Horn.

I've found this method to be useful: arguments have values which are
sets (of referents), and predicates are set-valued functions (in this
example, intersection with the complement of Marge).  This approach is
much more resistant to the siren song of "common sense"; if the result
is in extension when you think it ought to be packed up in a unit, then
you know there's an error somewhere, whereas once having decided that
extension is the right value, you know that you can't go back to the
unit, without additional verbiage.  

		-- jimc