[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
[sent 5 Nov]
pc:
> arosta@hidden.email writes:
> <<
>
> I don't know what terminology would be more familiar to you than
> "extensionally"
> and "intensionally" defined sets. An ext.-defined set is one defined
> by listing its
> members -- and they may have no uniquely common property other than their
> very membership of the set. An int-defined set is the set S such that
> every x is
> a member of S iff x has property P.
>
> >>
> OK, normal meaning. But then the problem is that e-sets (and le
> everythings pretty much) are defined extensionally and then given a
> label, which may or may not actually apply. That is, the members are
> picked first, then the "property". Or are you making a total change
> in this pattern here (it doesn't really seem so from other things you say)?
I've only very recently been thinking about this, but my current thinking
is that {le'i (su'o) broda} is unspecified about whether the set is
e-defined or i-defined, tho if it is i-defined, the defining property
is not specified (every member is a broda, but not every broda is necessarily
a member).
But, otoh, {le'i ro broda} would be an i-defined set, tho again with the
defining property unspecified. This is because cardinality ro allows for
cardinality 0. A 0-cardinality subset of lo'i broda cannot be defined
extensionally, so it must be defined intensionally.
--And.